*3.6. Trends*

The revision of the literature on service robots identifies diverse trends in the fields of healthcare, industry, home service, and multi-purpose indoor environments. Regarding the type of robots, trends in autonomous navigation [99,100], mobile robots [35], unmanned autonomous systems [101,102], and imitation learning systems [103,104] were identified.

Healthcare: The current global population trend has shown an increase in elderly persons, as well as an increase in more populated cities; therefore, there is a higher demand for healthcare services such as medicine and nursing [105]. According to Archibald and Barnard [105], three types of service robots oriented to health exist: doctor, nurse, and home health care robots. Regarding nursing robots, applications have been implemented for feeding assistance, automated soaping and showering for the elderly, robotic therapeutic companions, pharmaceutical transporters, pick-and-place patients in bed, and ambulation assistants [105]. A human action recognition algorithm using depth cameras, object detection, and human joint identification techniques supports patients with mild cognitive impairment (MCI) at their homes. Assistive living robots could be of use to MCI patients, as they can identify potential risks due to errors and avoid dangerous accidents [106].

Other studies have implemented physical exercise programs in elderly health centers with success. In this study case, 41 volunteers underwent a training program with a humanoid Aldebaran NAO robot and with a human coach [107]. The study results showed a good response towards the humanoid coach, at least for the training program. When analyzing a factual information task between robot and human, the human coach was more efficient. A biomechanical rehabilitation, one degree of freedom (DOF) robotic handle for post-stroke patients applies an adaptive reinforcement learning (RL) algorithm [108]. By constantly adapting the difficulty level of a virtual "nut-catching" game, depending on the skill level of volunteers, patients can learn at a faster pace while not losing their motivation.

A telemedicine robot (see Figure 10) was developed in [109], composed of a four-wheel base, a robotic arm, and a tablet, acting as the head of the robot. The robot also includes an array of ultrasound sensors and cameras. The robot allows automated navigation combining its sensors and actuators, including obstacle avoidance and object manipulation tasks (e.g., the floor and shelves). It implements routines such as TakeMedicine, WallFollowing, and Doorpassing. Other applications are also included, such as fetching, providing reminders, calendar, and interpersonal communication [109].

**Figure 10.** (**a**) Autonomous telemedicine robot system for assisted and independent living. (**b**) Simulated and experimental paths of an obstacle avoidance trial in a real-world scenario (center for the elderly), executed by the telemedicine robot developed in [109].

Industry: Different applications of service robots have been implemented for the industrial context. In [104], a machine learning technology enables a chatbot to provide support to customers in financial-product sales. Using robots to work continuously, text information from FAQs, call center response manuals, and office documents were used as input to a machine learning model to generate artificial conversation about bank services to interested customers. The use of unmanned aerial vehicles (UAV) has been reported in high risk tasks, such as transmission line inspection in China, Japan, Spain, and Britain [102]. By implementing UAV control and combining image processing and artificial intelligence, UAVs have performed autonomous inspection. These types of robots use a combination of visible light and thermal infrared sensing, as well as LiDAR technology [102]. Although UAVs are useful for safety purposes, manual inspection still outperforms in some scenarios, and it can also perform repairs, while UAV cannot.

Collaborative robots, usually robotic arms or semi-humanoid robots, represent an intermediate automation level between manual and fully automated manufacturing. In this approach, the robot acts as an assistant to a human during specific tasks. Vision-based collision prediction systems, capacitive sensing (skin detection), and safety design parameters, and routine instrument these types of robots [110]. Another typical application for industrial robots is object manipulation. An exciting study case is that of the first Amazon Picking Challenge [111], challenging 26 teams (primarily academics). The challenge involves designing autonomous robots to pick objects from a warehouse shelf. Objects of different shapes and sizes were used (Oreo cookies, an outlet protector, and a softcover book are some examples), and the teams used different approaches. Among the most

common solutions were suction actuators, 3D imaging sensing, and a geometrical and/or color recognition approach for feature selection.

Home service: Some application trends in home service are in education, entertainment, household, social interactions, gaming, security, and rehabilitation [112]. A home voice-activated semi-autonomous vehicle robot was implemented in [113]. It consisted of a modified lawnmower and an IoT control module through voice-activated Alexa commands. Humanoid robots can help in house chores, grasping and carrying objects, opening doors, and entertainment [114].

Multi-purpose indoor environments: Positioning systems using sensor fusion for indoor positioning tasks mainly uses ultrasonic sensors and information from radar and odometry [115]. A mobile robot using a robust convolutional neural network (CNN) algorithm for person identification, tracking, and locking followed the identified persons through different rooms, with great accuracy [116]. The development of a robotic waiter system integrates different autonomous navigation algorithms and sensing approaches, such as IMUs, odometry, SLAM, and adaptive Monte Carlo simulation [100]. Artificial vision methods identify tables in a restaurant as well as persons. Another approach of indoor positioning, Steady Delivery, makes use of sensor fusion, involving radar, ultrasonic sensors, and odometry [117]. Table 3 presents the current trends of service robots.


**Table 3.** Service robots' current trends.


**Table 3.** *Cont*.

#### **4. Conclusions**

From the presented review, the emerging status of service robots technology in the world and as a research area becomes more evident. This carries essential opportunities for early research, development, and investment in commercial technology as a strategic decision for long-term profit [118]. Despite the fact of the good will of robotics, some challenges are still present. Some of them are the lack of generalization and formalism in classifications and taxonomy [10], the current perceived utilitarian value [61], battery and autonomy modelling and estimation [119], ethics [36], and even design problems related to gender biases based on the occupation of the robot [120]. These challenges are opportunities for future research questions or different research groups. Moreover, a solid field of service robots is healthcare and cleaning robots. Consequently, with the COVID-19 pandemic, a push in these technologies was observed [37,90] and must be taken into account when performing research in this area. After the *accelerated development* in these technologies (caused by the COVID-19 spread) reaches a plateau, it may be an interesting research question to study the degradation or improvement in service robotrelated areas. As a final remark, despite the fact of current opportunities and observations in the field, estimations [50] point that, one way or another, service robots will undoubtedly be part of our daily life in the near future. This study will help researchers as it provides valuable information on recent developments of service robots. This work can serve as a starting point for researchers when studying this field. Moreover, we consider it helps to estimate the future value of the investment in service robot research and development. Consequently, the readers will contribute to this work by producing more studies and expanding the research area.

**Author Contributions:** Conceptualization, J.A.G.-A. and J.d.J.L.-S.; methodology, J.A.G.-A., R.O.-O., J.L.-I., K.L.R.-H. and M.A.R.-M.; investigation, J.A.G.-A., R.O.-O., J.L.-I., K.L.R.-H., and M.A.R.-M.; writing—original draft preparation, J.A.G.-A., R.O.-O., J.L.-I., K.L.R.-H. and M.A.R.-M.; writing review and editing, J.d.J.L.-S., M.A.R.-M., and R.A.R.-M.; visualization, J.A.G.-A., R.O.-O., J.L.-I., K.L.R.-H., and M.A.R.-M.; validation, J.d.J.L.-S., R.M.M. and R.A.R.-M.; supervision, J.d.J.L.-S., R.M.M. and R.A.R.-M.; project administration, J.d.J.L.-S., R.M.M. and R.A.R.-M.; funding acquisition, J.d.J.L.-S. and R.A.R.-M. All authors have read and agreed to the published version of the manuscript.

**Funding:** This project is funded by the Campus City initiative from Tecnologico de Monterrey. The APC was funded by Tecnologico de Monterrey.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** Authors would like to thank the support of Tecnologico de Monterrey through the Campus City initiative.

**Conflicts of Interest:** The authors declare no conflict of interest. The founders had no role in the writing of the manuscript.
