Industrial-Grade Edge Computing Device for Smart Furniture Products

Featured Application

Industrial IT communication technologies embedded in furniture enable manufacturers to expand market footprint and maintain sustainable production.

Abstract

As Internet of Things (IoT), communication convergence, and distributed computing are maturing, many new solutions are invented or derived every day, bringing new products and services to society. This paper elaborates on an industrial-grade information technology (IT) component upgrade to an existing ESP32 IoT device and the addition of new data communication layers to its capabilities. The aim of this research is to explore and propose new IT products to upgrade the first-generation smart furniture product. The proposed upgrades postulate a technological leap ahead in terms of (1) connectivity; (2) connection security; (3) product life-cycle management; and (4) integration (including sustainable product development). A distinct difference from the existing solution is presented in the aforementioned categories. The research points to a possible product that accommodates requirements and a niche of products that can be used, while eliminating other form factors. The process itself proves a point in the favor of the chosen IT product, providing confirmation of the hypotheses. Even though software and application security are not focal points of the discussion in this paper, these elements are generally presented as their presence is necessary in the modern convergence of communication methods.

1. Introduction

Although one might argue that the connected world has also brought significant downsides, mainly in social and (false) information sharing aspects of life [1], connectivity is indeed the most important driver of the Smart City, IoT, Industry 4.0, and associated concepts that bring new applications to everyday items and processes. Industry 4.0 is based on the intercommunication between previously unconnected items or items that were connected using incompatible systems, thus meaning that their connectivity was not in real time [2]. A new application layer on top of Industry 4.0 and IoT brings added value for making previously limited systems smarter. It enables a completely new overview of the entire system and provides a base for new Business Intelligence (BI) [3]. In industrial applications, this is easily perceivable through the comparison between the classic Supervisory Control and Data Acquisition (SCADA) applications and new Platform-as-a-Service (PaaS)-based applications. A similar example is the comparison between the classic programmable logic controller (PLC) and embedded computers with superior edge computing capabilities [4].

Smart city, smart living, and quality of life are just a few of several up-to-date topics covered in smart city aspirations, explained in detail in [5]. The Internet of People is a subset of the Internet of Everything or Internet of Things. It includes people communicating with items that did not previously exist or items that have been significantly improved (upgraded, retrofitted, or otherwise enriched—it is these processes that are crucial for defining the applicability of smart furniture in this paper). These items are intended to improve the quality of everyday life by providing information or enabling connectivity in areas where it did not previously exist. Edge computing, according to [4], involves a substantial amount of information being processed on “the edge” of a system, or as close as possible to an information source. As opposed to cloud-based systems, given the amount of information gathered and the requested latency (minimum possible), even with gigabit Internet access speeds, it is completely impractical to send vast data to the cloud for computing and send the results back. A natural extension to the IoT world is edge computing, with edge devices being capable of so much more than just sensors, logic controllers, and small-scale reduced instruction set computer (RISC)-based devices.

Modern hardware development has enabled the integration of significantly more powerful devices into increasingly small form factors. This enables integration at higher computing, connectivity, and application levels. Technology is present, and it is up to an individual business case to evaluate how integrated, robust, and powerful a device is supposed to be to serve its purpose.

This research suggests a (1) future-proof device with a (2) standardized form factor made from (3) industrial-grade components with an extended life cycle that is capable of withstanding various environmental and other usage conditions, creating a new connected smart furniture product that will make a significant leap forward for this particular project. This product is in line with existing industry standards. This step is an extension of an already manufactured (proof of concept) ESP32-based smart device made of commercial-grade hardware, with a basic connectivity set (with reduced security features), modest computing capabilities, and limited or no long-term replicability. The research results will provide investors (manufacturing companies) with a viable option to upgrade their product accordingly. More on the ESP32 device is presented and referenced in Section 3.

Smart furniture, as postulated by [6], can be best defined as “designed networked furniture that is equipped with an intelligent system or is a controller operated with the user’s data and energy sources”.

Although numerous different approaches have been used for this topic, the authors did not find clear evidence of focused approaches to industrial-grade design and connectivity integration into classic furniture pieces for everyday households. The keywords “Smart Home”, “IoT”, “Home IoT”, and “Smart Furniture” were used to search the SCOPUS, Web of Science, and IEEE Xplore databases. The remaining papers (that reference “smart furniture”) are commented on in Section 2. The researchers involved in this project are interested in upgrading their existing design and products to attract new customers and gain a new market niche that has been, so far, reserved for commercial home assistant digital stand-alone products.

On the other hand, smart furniture environments represent specialized ecosystems where intelligent furniture systems can operate at optimal functionality. These environments are characterized by a robust infrastructure of connectivity, sensing capabilities, and interactive technologies that enable furniture to respond dynamically to user needs and environmental conditions, and to control various connected devices.

One could easily argue that one of the defining characteristics of smart furniture environments is their comprehensive sensor (or, more in general, input) ecosystem. These environments typically integrate multiple sensing technologies, including photosensitive sensors, sound sensors, temperature sensors, and infrared sensors that allow furniture to detect or perceive the user’s actions, presence, touch, and other information (for inputs, that would be digital inputs or remote/wireless logical inputs), enabling the specific intelligent functions of automatic detection, automatic sensing and, in the end, the automated control of output/peripheral devices [7]. This sensor-rich environment enables the foundation for furniture that can adapt to changing conditions and user preferences without explicit commands.

Another defining characteristic, as presented by [7], is the presence of embedded processing systems (system on chip or similar) that function as the central control system of intelligent furniture. These environments require appropriate computational infrastructure to support the microcomputer processing system that is embedded in the furniture body. The embedded system serves as the integration point for software and hardware components, primarily composed of an embedded processor, operating system, electromechanical devices, and other supporting hardware and application software, as well as interconnected devices.

Connectivity infrastructure represents a third (vital) characteristic of smart furniture environments. These spaces require robust wireless networks (WiFi, Bluetooth, Zigbee, LoRa…) that facilitate communication between smart furniture (pieces) and connected devices. According to [7], users in these environments can use mobile phones, tablet computers, and other mobile terminals to control the specific functions of smart furniture through wireless networks or remote control. This interconnectedness enables seamless integration with broader smart home or even remote cloud ecosystems.

The applicability of smart furniture environments extends across various domains, as will be presented in more detail later in Section 2. In residential settings, these environments enhance everyday living through personalized comfort and convenience. One example would be a smart bed that integrates a computer, TV, and game system, turning the bed into a powerful entertainment system. In office environments, smart furniture supports productivity and wellness, exemplified by solutions like the Autonomous Desk, which adjusts its height and reminds users when the user’s standing or sitting time exceeds a reasonable range of health [7].

Healthcare facilities represent another application area where intelligent furniture can support patient monitoring and care. Educational institutions can leverage smart furniture environments to enhance collaborative learning and resource management. In hospitality settings, these environments can elevate patient experiences through personalized accommodations and interactive features.

Smart furniture environments also present specific challenges, including power management considerations, privacy concerns, and interoperability issues. As noted in research on smart homes, these environments must balance technological capabilities with user needs to provide a comfortable, convenient, safe, and joyful life through the managing of various technologies. These environments must implement thoughtful design approaches that prioritize user experience alongside technical functionality, as demonstrated in [8].

This paper’s structure is as follows:

(1)

Introduction: General terms associated with the topic and research hypotheses; declaration of items.

(2)

Related work: An overview of the existing research with common keywords; validation of interest.

(3)

Authors’ contributions—existing product characteristics and future product requirements/criteria: Lay-out of the existing product with its limitations and shortcomings. A list of user-defined criteria to steer research and define components for the new product; validation criteria.

(4)

Solution/candidate product.

(5)

Discussion: Validation of the existing product and new product analysis.

(6)

Conclusion: Recapitulation of the results and a final product suggestion.

2. Related Work

The term smart furniture is present in many different applications and some authors have gone to great lengths to identify and define it [6]. From a wider perspective, smart furniture is found in many different aspects of life (not just homeware), yet most products are connected or related to the smart city umbrella. For example, Ref. [9] expanded this term into items publicly available on the streets of cities that are placed and applied in order to improve quality of life and to specially support some population segments, such as the elderly population. A similar example is presented in [10], where certain items (kiosks) were identified as a point of contact. The items presented in these applications mostly focus on points of information and human–machine interface devices. Another example, presented in [10], is a digital smart table conceived and defined as a helpful digital interactive tool for fast response teams in emergency operation centers. The digital table is built around a digital imaging device and sensor network that is intended to identify human movement, hence facilitating human interaction with the system. There are several examples of medical applications for smart furniture as well. Ref. [11] explains a specially created chair designed to take heart measurements that can later be processed and interpreted. A case study for telehealthcare devices embedded into hospital furniture is presented in [12]. Yet, none of these applications focus on the manufacturing of general-purpose enhanced home smart furniture that would serve as an IT hub, even though some items are enhanced in connectivity.

Perhaps the two most prototypical examples of using smart furniture reflect either of the following:

(a)

The smarter way of designing classic furniture to save space and make more efficient furniture by using a smart design or innovative materials, or their combinations;

(b)

Some sort of integrated human–machine interface device (in many cases, optical touch-screen devices or similar), as presented by [13].

Since the term smart furniture is rather broad, the need to focus the research was imperative. One such overview was presented in [14], following modern concepts of interconnected devices. However, this market niche and related products have shifted partially to wearable segments, and only some (mainly infrastructural) products could still be considered as smart furniture. A similar presentation was made in [15] in regard to researching the points where the smart design of furniture meets digital and connected furniture and its application.

Continuing to focus on the desired smart furniture niche, Ref. [16] declares and evaluates several main types of smart or connected furniture. These are a pole type, lamp type, mirror type, and a message board type, with different types of middleware. The presented furniture types are significantly more in line with the furniture presented in this paper, yet no concise approach to edge computing or an IT hub is given in any of these sources. Following the already mentioned example of medical applications [11,12], an interesting discussion is presented in [17], arguing that the intelligent usage of embedded sensors in a couch could lead to better sleep apnea detection. An overhaul view of the interdisciplinary approach to the consideration of furniture design and the smart furniture concept was given by [18,19], emphasizing elderly-focused commercial ICT products. These papers focus on a very narrow niche. Having researched the above, the authors wanted to confirm that the market maturity is indeed at an appropriate level before continuing. A digression in research was made as follows:

According to the European Patent Office [20], a significant number of reported patents are identified in the period of 2010 to 2021 (2022) for the keywords “Smart Home” and “IoT”. The cumulative patent count is 180.589 and 386.285, averaging more than 40 patents per day for “Smart Home” and just below 90 patents per day for “IoT”. Since these keywords can be associated with many other items not related to home Internet of Things or smart furniture, the authors searched for patents mentioning the keywords “Home IoT”, which returned a less extensive result list, averaging 0.39 patents per day, yet with a very indicative distribution, with 1 patent per year in 2010 and more than 1 patent per day in 2020, showing a significant growth rate. The additional keyword “Smart Furniture” was used to better understand the results. More general keyword usage gave us more results (two-to-three-fold) in relation to “Home IoT”.

Table 1. Number of published patents per year according to keywords.

Year	Smart Home	IoT	Home IoT	Smart Furniture
2010	500	781	1	2
2011	618	1.200	3	11
2012	1.092	1.826	9	9
2013	1.627	2.507	18	28
2014	3.864	4.212	41	53
2015	10.552	7.559	39	193
2016	16.417	15.284	116	285
2017	17.909	27.719	209	368
2018	24.159	55.747	252	853
2019	26.630	75.442	340	709
2020	34.871	92.802	380	821
2021	42.350	101.206	314	1.092
Cumulative	180.589	386.285	1.722	4.424

lists the aforementioned data in raw numbers and

Table 2. Yearly growth rate of published patents according to keywords.

Year	Smart Home	IoT	Home IoT	Smart Furniture
2010	-	-	-	-
2011	23.60%	53.65%	200.00%	450.00%
2012	76.70%	52.17%	200.00%	−18.18%
2013	48.99%	37.29%	100.00%	211.11%
2014	137.49%	68.01%	127.78%	89.29%
2015	173.08%	79.46%	−4.88%	264.15%
2016	55.58%	102.20%	197.44%	47.67%
2017	9.09%	81.36%	80.17%	29.12%
2018	34.90%	101.11%	20.57%	131.79%
2019	10.23%	35.33%	34.92%	−16.88%
2020	30.95%	23.01%	11.76%	15.80%
2021	21.45%	9.06%	−17.37%	33.01%

represents the yearly growth in specific numbers of published patents.

In conclusion, the data clearly show that the Internet of Things is associated with a great number of patents in 2021 (101.206), and the term smart home is also on the rise. The two can be considered to be more general terms that will result in connectivity being a part of more and more items in our lives.

Home IoT, on the other hand, is a much narrower term that is closely associated with the household Internet of Things. This term demonstrates strong yearly growth rates. Smart furniture is yet another more general term, but still manages to show an interesting growth rate in a parallel time frame.

The second criterion to establish the market trend was to research strings used by possible users via the Google Trends search engine.

Table 3. Heatmap presenting values of Google Trends’ “Home IoT” keyword usage in period 2010–2020.

	Year
	2010	2011	2012	2013	2014	2015	2016	2017	2018	2019	2020
Jan	55	0	0	0	0	27	42	45	29	41	81
Feb	0	0	18	0	0	0	46	27	64	84	70
Mar	0	0	0	0	0	17	26	15	85	76	71
Apr	54	0	0	0	0	0	0	75	89	56	68
May	0	0	0	0	0	17	41	81	79	50	59
Jun	0	0	0	0	0	0	15	50	44	25	44
Jul	0	0	0	0	0	33	0	65	66	62	52
Aug	0	24	0	0	27	15	0	55	26	53	79
Sep	0	0	27	0	29	11	40	57	67	43	50
Oct	0	0	0	0	0	13	40	79	87	75	52
Nov	0	0	49	0	24	33	25	47	100	55	56
Dec	0	22	0	0	0	26	39	70	66	45	45

shows a relevant trend (shown as a percentual relative number in relation to the maximum recorded search string detected input—date Nov 2018) in the form of a heat map/table. The green values represent lower hit cases, whereas warmer and red values represent higher hit cases.

The above data show an increase in users’ interest in Home IoT starting from 2016, which shows a correlation with the patent count in the same time frame.

Taking the aforesaid into account, the authors conclude that there is sufficient market activity to encompass and consume possible future products like smart furniture that would be a part of the home IoT ecosystem.

Despite the mentioned stand-alone commercial products that serve as multimedia hubs or smart home centers, a clear differentiation criterion is made between customer-grade commercial products and integrated furniture based on solid industrial-grade design and components. The major differences between the existing solution and the product suggested by the authors are presented in a table on lines 461–464 in Section 5.

3. Existing Product and Future Product Requirements/Criteria

An ideal electrical or electronic device can be connected to an external power source all the time and is not affected by the climate and mechanical restrictions of the place it is installed in. A real device needs to be thoughtfully conceived, designed, and constructed to meet all restrictions imposed by the design and other elements. As presented in [15], a need for design changes and possible standardization for the acceptance of information communication technology (ICT) components into furniture units has been brought to researchers’ attention. This standardization will be one of the key points in this paper, as it enables manufacturers to rely on a single mechanical design.

In papers [13,21], the authors propose adapting a product from an existing furniture manufacturer to include information technology (IT) elements, making the furniture smart in the way that it uses numerous functionalities to reduce human effort and increase quality of life by offering new IT and convenient functionalities. The foundation for this smart furniture is contextually connected to the smart home, smart device, and smart environment concepts. The authors of [14,21] identified some of the key characteristics for smart furniture to be accepted as a new product. Some of these characteristics correspond to the restrictions stipulated in the introduction and can help steer future designers in creating a more refined product. These characteristics were functionality, design, safety in use, customization, and structural design. The proposed products were various systems on a chip with elaborated communication modules to assure affordable connectivity solutions and integration with personal mobile devices. The key requirements were simplicity and a product at a proof-of-concept level.

The ESP32 platform was chosen for initial prototyping due its following qualities:

-

Versatile Connectivity Options:

ESP32 is equipped with integrated Wi-Fi and Bluetooth/Bluetooth Low Energy (BLE) capabilities, making it ideal for seamless communication between furniture components and external devices. This hybrid connectivity allows ESP32 to act as a standalone system or as a secondary device interfacing with other systems via SPI, SDIO, I2C, or UART interfaces [22].

-

Cost-Effectiveness:

The affordability of ESP32 makes it an attractive choice for proof-of-concept development. Ready-to-use ESP32 development boards are available at approximately USD 6 per unit, significantly cheaper than alternatives like Nordic Semiconductor’s nRF52840 (USD 20) or STM32 microcontrollers (USD 5–USD 30) [23] as shown in

Table 4. Cost comparison.

Chip	Price Range	Features
ESP32	USD 3–10	Wi-Fi + Bluetooth/BLE, Dual Core, 500 kb SRAM
STM32	USD 5–30	ARM Cortex-M Series, Single/Dual Core 192 kB SRAM
nRF52840	USD 7–20	Bluetooth/BLE + NFC, Single Core, 256 kB RAM

.

-

Energy Efficiency:

ESP32 is engineered for ultra-low power consumption, making it suitable for battery-operated smart furniture designs. It features fine-grained clock gating, dynamic power scaling, and multiple sleep modes (light sleep, deep sleep, hibernation) [D].

Power Consumption Details:

Active Mode: Consumes approximately 160 mA during full operation.

Deep Sleep Mode: Reduces power consumption to as low as 10 µA by shutting down most peripherals while keeping essential components like the Real-Time Clock (RTC) active.

Hibernation Mode: Further minimizes power usage by disabling all components except wake-up triggers.

These energy-saving features allow smart furniture devices to operate efficiently over extended periods without frequent battery replacements or recharging.

While suitable for prototyping, ESP32’s limitations in processing power and scalability necessitate the upgrade to industrial-grade devices described in this paper.

The authors defined and pointed to a commercially available product to be installed into classic furniture to achieve the aforementioned functionalities. Such products are not manufactured from higher-quality-grade components, but from classic- or commercial-grade components. These products are not necessarily standardized in size or form factor, nor are supported by a manufacturer in the long term, with backward mechanical and general system compatibility. Hence, the proposed product poses challenges for system (furniture) integrators who plan on using it in their products for a long time.

The authors propose an industrial-grade product with a defined form factor that demonstrates an improvement in all important segments of implementation.

The requirements were also given as inputs. The requirements given to the authors for the next-generation smart furniture product were as follows:

• Minimal physical footprint;
• Standardized form factor;
• Increased computing power and memory while maintaining low consumption;
• Passively cooled and fan-less device with no moving parts;
• x86 compatibility (to be able to run modern architecture software);
• Connectivity expandability (at least three different methods + two Ethernet ports);
• Integrated system on a chip (SoC) to avoid practical procurement issues and thermal dissipation issues;
• Compatibility with current major cloud applications/providers;
• Ability to run modern communication security protocols.

The listed specifications point to a device with discrete physical dimensions, a standardized form factor, per-design interoperability, reduced thermal dissipation combined with fan-less operation, increased computing power with compatibility with modern computing architecture (x86 and 64 bit applications), and the ability to facilitate multimodal standardized communication technologies, such as Wi-Fi, Bluetooth, Mash Bluetooth, Zigbee, and the like. Integrated SoC was preferred to maintain standardized configuration and keep the thermal dissipation of the entire system predictable. Even though these requirements fit a particular use case, the authors suggest that any similar case might benefit from the same or similar criteria.

The authors propose an extension to a previously presented concept (existing product based on ESP32 board) by introducing a more powerful (both in terms of computing and storage power) fan-less embedded edge device instead of the described system on a chip, which would be able to perform far more simultaneous functionalities (including enhanced security features) and would be future-proof in the way that it enables a manufacturer to maintain mechanical and thermal dissipation design regardless of the current device being installed. Several of the most used form factors are presented in [24].

A device must operate both on a local network either as a client or a server (or both) and be connected to the Internet for an additional level of control and data access. Additional ports and connection points are needed to establish wired communication with devices that are controlled by the edge device. Functional elements of an example IoT hardware and software concept are described in [25]. The edge device must be able to communicate with several sensors or other devices via integrated communication protocols (such as Zigbee, Bluetooth, or Wi-Fi).

Aside from hardware and environmental requirements, a software platform should be connected to facilitate easier access to the device and its functionality, as well as to control connected devices. This paper does not discuss or prefer a particular software solution or platform, as the proposed device should be compatible with several modern operating systems. Overall, stable, secure, and functional communication must be achieved, emphasizing software, connection, and platform software security. This, again, is a different sub-topic and is not the focal point of this paper.

The authors declare the following contributions through a systematic hardware selection methodology.

This paper makes several significant contributions to the field of intelligent furniture design, particularly through the development of a systematic approach to hardware selection that optimizes selection through the given requirements. Unlike previous research that may focus primarily on conceptual aspects, our work addresses the critical hardware implementation challenges that determine practical viability.

A primary contribution is the development of a comprehensive hardware evaluation framework specifically tailored for intelligent furniture applications. While intelligent furniture comprises hardware facilities with intelligent functions and software facilities with intelligent programming systems, the selection of appropriate hardware components directly impacts all metrics. Our methodology is similar to the systematic evaluation approach demonstrated in [8]. This paper introduces a multi-criteria decision-making process for hardware component selection that considers essential metrics, including form factor, thermal design power, x86 compatibility, expandability, and communication functionality, as well as connectivity. This approach recognizes that intelligent furniture hardware must balance computing power, TDP, and expandability with physical constraints, as furniture pieces must maintain their primary functionality while incorporating technology.

By providing this systematic approach to hardware selection, this paper advances intelligent furniture design beyond conceptual models to practical implementation guidance, addressing a gap in the current literature that often emphasizes design principles or software or AI perspectives, without sufficient attention given to hardware considerations that are basic for any smart software add-on.

4. Solution/Candidate Product

Since the project goal is to provide a product suitable for the sustainable development and life-cycle management of final houseware products with an emphasis on standardized components, using products from the standardized industrial form factor is a must-have requirement, as the manufacturer needs to maintain mechanical design and thermal solutions for as long as possible and to have this candidate product included in as many separate products as is practical. For the same reason, the commercial-grade solutions that were present in the proof-of-concept (first-generation smart furniture based on ESP32 device) products are not accepted, as their life cycle and the lack of an industry standard contribute to risk factors.

Below is a

Table 5. Initial candidates and their characteristics.

Form Factor	Dim. [mm]	Z-Height [mm]	Max TDP [W]	Expansion Support	SOC	Key Features
1.8″ SBC	100 × 65	10–15	5	-	Ready	Ultra-compact design, minimal I/O footprint
2.5″ SBC	100 × 70	15	8	MiniPCIe	Ready	Compact storage-optimized layout, moderate expansion
3.5″ SBC	146 × 102	20	10	MiniPCIe+M.-2	Ready	Balanced size, standardized mounting, optimized IoT protocol support
Nano-ITX	120 × 120	20–25	15	MiniPCIe+M.-2	Support	Small footprint with good expansion options, suitable for embedded systems
Mini-ITX	170 × 170	40–45	35	MiniPCIe+M.-2	Support	Desktop-grade performance, PCIe expansion
COM Express	125 × 95	20–25	45	M.2	Support	Modular design, high-performance computing, supports multiple CPU architectures
Q7	70 × 70	12–15	12	M.2	Ready	Compact modular design, optimized for low-power ARM and x86 SoCs
PC/104	90 × 96	15	5	-	Ready	Rugged FPGA security architecture, industrial-grade durability
JEDEC B111A	132 × 77	10	15	-	Ready	Stress-optimized design, enhanced solder joint reliability
Timepix4 c.	80 × 300	25	20	-	-	High-speed data acquisition (160 Gbps), elongated form factor
Compact PCI	100 × 160	30	25	MiniPCIe	Support	Backplane connectivity, hot-swappable modules

of all the industrial form factors that were initially considered, with their main characteristics and features/restrictions.

All industry formats in Com Express and Q7 form factors were dismissed as they were too small and too compact to accommodate the needed expansions and connectors, and they lacked the needed power and (mostly) x86 compatibility. All ATX form factors were dismissed due to oversized dimensions, power and cooling requirements, and mostly unavailable SoC solutions. PC/104, JEDEC B111A, and Timepix4 were dismissed due to a lack of expandability. Compact PCI was not a form factor that was favorable due to its requirement to have an active board as an additional piece of equipment needed. The remaining standardized form factors were in the ITX segment. The candidates are presented in

Table 6. Potential candidates and their characteristics.

	Mini ITX	3.5″ Board	Nano-ITX	2.5″ Board	1.8″ Board
Board size [mm]	170 × 170	146 × 104	120 × 120	100 × 72	85 × 56
x86 compatible	YES	YES	YES	YES	YES
Multi display	4	3	3	2–3	NO
SoC integrated	N/A	YES	YES	N/A	YES
Expansions	3+ (Full size)	2–3	2	1 (up to 2)	1
LAN	2	2–4	2	1–2	1
Flat design	NO	YES	NO	YES	YES

, which shows the selected characteristics of the different-size ITX form factor boards [22].

Considering the defined requirements, with a special emphasis on physical dimensions and thermal solutions (related to the thermal dissipation characteristics of the central processing unit), the Mini-ITX form factor was eliminated for being oversized in all dimensions and lacking SoC integration in most instances (although some products have SoC implementation), as well as lacking Mini-PCI or M.2 form factor communication expansion slots. Moreover, 2.5″ and 1.8″ boards were dismissed for their general lack of expandability, and Nano-ITX options were dismissed due to their high z-height dimension and lack of expansions.

The selected single-board PC (SBPC) belongs to the remaining group of 3.5″ boards that met all of the required elements (both physical and logical).

The board itself is less than 15 × 10 cm in size and less than 2 cm in height (not breaching the height of the needed connectors). The thermal solution and chassis can be adapted to serve installation restrictions. A power supply is needed (DC, 12–24V).

An example of the product (top side) is shown in Figure 1 [26,27].

The basic IT specifications of this device are as follows:

Intel Celeron N6210 processor (dual-core processor with a maximum clock rate of 2.6 GHz). The thermal design power is 6.5 W, assuring the minimum thermal dissipation and influence on the chassis design and system performance. It is 64 bit, with smart connectivity capabilities capable of housing virtual machines and applications such as additional security software. An example of the device can be found in [23]:

• Random access DDR4 memory: 4 GB;
• CFast-type storage module connected to SATA port: 16 GB or 32 GB;
• Up to four serial ports for legacy communication (two of them being RS-422/485-capable);
• Gigabit Ethernet RJ-45 port + 2.5 Gbit RJ-45 port;
• DI/O—eight channels;
• Temperature working range: −20 up to +70 °C, with 10 to 95% relative humidity.

The other/additional items visible in Figure 1 are the following: internal LVDS and USB connection points; internal COM points for previously presented COM ports; DC-in point; SystemOnChip; eD point; external HDMI, DP, USB, and Ethernet ports. Expansions to this basic board are possible in the form of additional communication modules that can be fit to Mini-PCIe or other onboard M.2 slots. These boards are needed in order to expand communication capabilities, as these were some of the requirements.

Table 7. Potential candidates and their characteristics.

	Mini ITX	3.5″ Board	Nano-ITX	2.5″ Board	1.8″ Board
Minimal physical footprint	NO	YES	NO	YES	YES
Standardized form factor	YES	YES	YES	YES	YES
Increased computing power and memory while maintaining low consumption	YES	YES	YES	YES	YES
Passively cooled and fan-less device with no moving parts	NO	YES	YES	YES	YES
x86 compatibility (to be able to run modern architecture software)	YES	YES	YES	YES	YES
Connectivity expandability (at least 3 different methods + 2 Ethernet ports)	YES	YES	NO	NO	NO
Integrated SoC to avoid practical procurement issues and thermal dissipation issues	N/A	YES	YES	N/A	YES
Compatibility with current cloud applications/providers	YES	YES	YES	YES	YES
Minimal physical footprint	NO	YES	NO	YES	YES
Standardized form factor	YES	YES	YES	YES	YES
Increased computing power and memory while maintaining low consumption	YES	YES	YES	YES	YES
Passively cooled and fan-less device with no moving parts	NO	YES	YES	YES	YES
x86 compatibility (to be able to run modern architecture software)	YES	YES	YES	YES	YES
Connectivity expandability (at least 3 different methods + 2 Ethernet ports)	YES	YES	NO	NO	NO
Integrated SoC to avoid practical procurement issues and thermal dissipation issues	N/A	YES	YES	N/A	YES

lists the selected product conformance with the requested specification.

Single-board computers (SBCs) have emerged as essential components in Internet of Things (IoT) applications due to their integrated design, combining processing, memory, and I/O capabilities on a single circuit board. This analysis examines industrial single-board PC form factors through four critical criteria: physical form factor, TDP, connectivity and expansion options, and x86 compatibility, with particular focus on the 3.5″ SBPC form factor’s suitability for connected furniture in home environments. As furniture becomes increasingly integrated with smart home ecosystems, selecting appropriate computing platforms becomes crucial for successful implementation and user adoption. This research synthesizes academic findings to determine whether the 3.5″ SBPC form factor represents an optimal balance of these criteria for furniture-embedded computing systems.

Industrial SBCs come in various standardized dimensions that impact their integration potential in furniture applications. According to [28], SBCs are evaluated based on certain characteristics, including physical dimensions, cost, processing capacity, and power consumption. The 3.5″ form factor (approximately 102 mm × 146 mm) represents a balance between computational capability and physical size that is advantageous for furniture integration. In [29], the authors’ work on portable computing solutions utilized a similar form factor (3.5 by 5 inch) that proved effective for space-constrained applications. The standardized mounting holes facilitate installation across various furniture designs while providing sufficient space for essential components without compromising furniture aesthetics or structural integrity and allowing for future upgrades or module replacements, while maintaining the form factor for backward compatibility.

The described device is capable of serving as the edge device for multipurpose communication systems, such as the following:

• Bluetooth HMI devices (similar to Bluetooth speakers);
• Bluetooth concentrators/servers (concentrators for smart home devices that connect through a Bluetooth protocol, like smart bulbs, refrigerators, or similar devices);
• ZigBee concentrators/servers (concentrators for other smart home devices that utilize a ZigBee protocol, such as smart (door, window, water leakage, gas detection, movement) sensors or different smart wearable devices that are a part of Internet of People concepts);
• Wi-Fi concentrators/servers/clients/hot spots/hubs (as concentration points for home surveillance cameras, other security devices, or devices connected using Wi-Fi protocols, like home robots and similar products);
• WWAN communication modules for independent connectivity to the Internet, regardless of the local infrastructure.

Such devices can be interconnected throughout the home-making mash capable network for redundant connectivity.

Figure 2 shows an example of the placement of a smart furniture edge device(s) as a hub in a smart home, as well as a networked living environment.

The device can depend on its own computing power and installed applications for managing tasks, or can interchange requests with connected personal mobile devices (mobile phones, tablets, or similar devices that already have AI assistants installed) to provide the best possible reply for all tasks and inquiries made by people/users/AI assistants.

Some of the mentioned functionalities can be more or less important to certain operating systems (e.g., AI assistants are more present by default in Android/Apple OS-based devices, whereas Windows-based devices need additional application upgrades to support both functionality and interconnectivity). On the other hand, Windows-based devices can produce more system stability and watchdog functionality. One might argue that Android OS is more stable compared to Windows OS due to the fact that it is much easier in Windows OS to interfere with the basic functions of the OS itself, yet the security features of Android are permission-declaration-based [30], so Windows OS can compensate for notable differences described above.

The most usable functionalities of the device can be allocated to the vendor’s cloud platform for easier access from a remote location. Also, this platform may serve as a software/application connectivity point for accessing all devices that are connected to the edge device.

An example of such an online platform can be found in the Advantech WISE-PaaS [31] product series, smart platform [32], or a similar online platform based on a Cumulocity IoT platform.

Even though this platform adds value to the performance and application of the edge device, the device can serve both as a local server and a client. In the case of Internet connectivity interruption, when on premise, the device can act as a server and maintain all functionalities locally, with communication to users (people with personal devices or ones communicating with the device via voice HMI) maintained via local WLAN or Bluetooth connection. WWAN-equipped devices can establish a back-up connection to the Internet regardless of the local infrastructure.

As the proposed edge device can be utilized as a multipurpose computing device, additional functionalities and applications can be added or uploaded remotely through the process of regular updates, ranging from safety features to ones making lives easier. The implementation of such a device can turn furniture into a central point (a hub) for maintaining a smart home regardless of the installed equipment, and can be vendor-agnostic as long as the installed third-party equipment follows open standards and is capable of interconnecting with similar equipment, or has open application programming interfaces (APIs).

The edge device can also accept methods of communication that are not specified in this paper simply by replacing Mini-PCIe/M.2-based communication modules with different ones or connecting other communication modules to serial ports.

Although the suggested device is based on a Windows 10 Enterprise LTSC operating system, a similar device can be made available using Linux, or specifically Android-compatible/based devices. Due to the specific requirements of Linux or Android OS, some functionalities or expansion capabilities may be reduced (e.g., the number of Mini-PCIe/M.2 slots or other ports).

Given the popularity of Android OS, some manufacturers or developers might prefer this OS to Windows-based devices.

4.1. Security Aspects

When creating any modern information technology device that utilizes external communication, an ever-growing and important aspect is information and communication security and safety. To maintain a satisfactory level of communication security and cyber security in general, software and platform designers must ensure that selected methods and protocols are adequate and are able to perform well under various reasonable circumstances that can arise during everyday use.

Software items in this regard can be divided into several categories. These categories are operating system and firmware, local drivers and software, and platform software and communication.

4.1.1. Operating System and Firmware

The selection of firmware and the operating system can be crucial and will steer an entire project. There are positive and negative aspects to selecting various operating systems. Embedded devices and their applications prefer using stable long-term support operating systems that do not force any specific regular updates or maintenance requests (unless pushed by the manufacturer for safe upgrading purposes). On the other hand, some operating systems are less flexible than others, so choosing an operating system can be a task that includes many compromises. Firmware is provided by a board manufacturer and generally supports all needed basic functionalities and APIs for the ports provided.

4.1.2. Local Drivers

Local drivers can result in weak security. If used improperly, wrong or third-party drivers can add uncertainties or unstable work, or even a backdoor entrance into some devices. For this reason, drivers must always be trusted and certified to work with a specific device. A preferred method is to have drivers under their own control when creating embedded devices.

4.1.3. Platform Software

Platform software (or cloud software) is a software application or a set of applications that is run at a separate datacenter location and is available and accessible via the Internet regardless of the availability of the local (edge) device. This platform is highly exposed to public access, and will be specially considered regarding communication and cyber security. Platform software is often proofed by penetration tests to assure the ruggedness and safety of the system, data, and communication channels. The system must be resilient to outside attacks and must always provide functionality. Data and especially user data must be kept safe from any leakage. User data that are associated with the edge device are completely sensitive.

4.1.4. Communication Channel(s)

A communication channel must always be protected either by using a Virtual Private Network (VPN) connection or some other similar method of secure connection. Some networks like the Narrow-Band IoT cellular network can provide safe communication channels that are not interconnected with the “Internet” (utilizing private the Access Point Name (APN) WWAN method) and can assure separate data connection to a desired datacenter or a private Internet Protocol (IP) address. Assuring safe and secure communication is paramount in creating a robust and safe system.

Some communication devices/modules are certified for a secure connection to certain platform services like Microsoft Azure Sphere [33] and similar services. Such devices embed needed protocols and communication methods to facilitate the easier establishment of secure connections on the WAN side of the system.

On the internal side, there are local connections present, mostly with sensors or user devices. These connections are facilitated using Wireless Local Area Network (WLAN) [34], Bluetooth (BT) [35], Zigbee [36], or similar protocols. All of these communication methods have their vulnerabilities that must be addressed when creating a local system around the edge device.

5. Discussion

A manufacturer interested in creating a new product, attracting new customers, and gaining a broader market reach has expressed its interest in researching new options and opportunities on how to upgrade their initial proof-of-concept-level smart furniture project. The initial project was based on proof-of-concept ESP32 hardware with a minimalistic software solution and reduced connectivity and operability functionalities.

The proposed upgrades enable planning new variants of their furniture, utilizing industrial-grade devices freely available on the market for easier design, planning, manufacturing, maintenance, and marketing. The proposed device is standardized in size and requirements, future-proof in its capabilities, and enables the creation and maintenance of elaborate firmware, software, and platform (application) software, adding extra value to the original product. Market and industry trends show significant market maturity for stakeholders to continue with this proposal.

Even though there are commercially available products similar to multimedia hubs and the like, the hereby proposed product is an industrial-grade product and rugged, assuring the proper functionality of the system in various scenarios. Decision-makers have an option to choose their path when it comes to selecting an operating system and a platform.

Since the product is standard in size, there are no-to-minimal risks in maintaining the long-term mechanical and thermal dissipation design. The additional items required for this product to run are a minimalistic chassis, power supply, and external antennas (if desired). Additional expansions, such as local speakers, lights, and wireless charging cradles, can be installed. A comparison between the existing hardware and the new candidate is shown in the table below (

Table 8. Comparison of the existing and new product.

	Existing ESP32-Based Device	New 3.5″ Industrial Board Device
Minimal physical footprint	YES	YES
Standardized form factor	NO	YES
Increased computing power and memory	NO	YES
Passively cooled and fan-less	YES	YES
x86 compatibility	NO	YES
Connectivity expandability	NO	YES
Integrated system on chip	YES	YES
Compatibility with current cloud providers	NO	YES
Higher security applications	NO	YES
Easily integrated	NO	YES
Scalable	NO	YES

).

The true value of this product (hardware-wise) is in its standard format, ruggedness, expandability, stability, interconnectivity, and general ability to serve as a central point of a smart home, connecting all items present in the household. Software-wise, it uses platform-agnostic hardware that enables the optimum choice of an operating system and associated applications. Energy-footprint-wise, this product can replace several other/parallel ones and enable the customer to save on electricity consumption. The solution is vendor-agnostic and can serve as a future-proof concept, as the furniture manufacturer can easily select different vendors to supply the standardized industrial IT product by selecting the favorable specification among those offered in the given form factor.

Security aspects can be additionally underlined with the proper choice of adequate communication modules that are ready to accept certified/secure communication protocols.

6. Conclusions

In this paper, the authors present a highly integrated industrial-grade edge device that is suitable to be implemented in the smart furniture project, providing novel and previously unused functionalities compared to classic furniture and the first-generation smart product while maintaining a standardized mechanical and thermal design, enabling a sustainable and future-proof solution. The described furniture products can be fitted as nightstands, club tables, integrated into mirrors, or adapted into wall-mounted items that provide all of the presented functionalities. Moreover, more may be added through visual HMI functions. The standardized form factor for ICT equipment can provide a sustainable framework for furniture designers and manufacturers, providing an industry-stable form factor and assuring continuity in the chassis, cooling design, and development.

A connected home is made possible by smart furniture that serves as a point of interest, a human–machine interface, and a hub for various devices that are either already present in the home or will be purchased because of the possibilities that existing smart furniture infrastructure can provide.

The open concept and interoperability allow the manufacturer to design and equip the product with less or more elaborate IT hubs (edge devices), or to adapt it to a particular business case by using more or less rugged/durable products, or a more or less powerful device capable of performing different levels of calculations. Aside from the physical characteristics and the form factor, the main added value in this product is the software application platform that can connect and utilize the hardware in many different ways and applications.

The development of such an application platform (both local to act as a local server and online as a cloud platform) is possible utilizing one of many already available resources. Vendors that supply hardware solutions that can be used as edge devices in this project also provide out-of-the-box Platform-as-a-Service (PaaS) solutions and applications that can be used and customized in an Original Equipment Manufacturer (OEM) or Original Design Manufacturer (ODM) project. The technology that supports the Internet of Things is present. It is down to the particularities of a business case to shape and fine-tune possible projects and products. The market is mature and ready.