Performance Characteristics of Newly Developed Real-Time Wave Measurement Buoy Using the Variometric Approach

Abstract

Accurate measurement of ocean wave parameters is critical for applications including ocean modeling, coastal engineering, and disaster management. This article introduces a novel global navigation satellite system (GNSS) drifting buoy for surface wave measurements that addresses the challenges of performing real-time, high-precision measurements and realizing cost-effective large-scale deployment. Unlike traditional approaches, this buoy uses the kinematic extension of the variometric approach for displacement analysis stand-alone engine (Kin-VADASE) velocity measurement method, thus eliminating the need for additional high-precision measurement units and an expensive complement of satellite orbital products. Through testing in the South China Sea and Laoshan Bay, the results showed good consistency in significant wave height and main wave direction between the novel buoy and a Datawell DWR-G4, even under mild wind and wave conditions. However, wave mean period disparities were observed partially because of sampling frequency differences. To validate this idea, we used Joint North Sea Wave Project (Jonswap) spectral waves as input signals, the bias characteristics of the mean periods of the spectral calculations were compared under conditions of identical input signals and gradient-distributed wind speeds. Results showed an average difference of 0.28 s between the sampling frequencies of 1.28 Hz and 5 Hz. The consequence that high-frequency signals have considerable effects on the mean wave period calculations indicates the necessity of the buoy’s high-frequency operation mode. This GNSS drifting buoy offers a cost-effective, globally deployable solution for ocean wave measurement. Its potential for large-scale networked ocean wave observation makes it a valuable oceanic research and monitoring instrument.

1. Introduction

Accurate measurement of ocean wave parameters, including wave height, wave period, and mean wave direction, is of paramount importance in various fields, including numerical ocean circulation modeling, ocean-atmosphere interactions, coastal infrastructure engineering, and tsunami early warning systems [1]. Recent studies further emphasize the critical role of waves in climate dynamics, illustrating how wave patterns directly influence climate variability and atmospheric conditions [2]. Research into surface waves contributes to a better understanding of climate system models [3], heat flux [4], and wave-turbulence interactions [5,6], which are pivotal in forecasting and managing marine and atmospheric environments. Furthermore, investigations into the impacts of waves on surface mixing processes [7] have highlighted their significance in oceanic surface layer vertical mixing. Acquisition of extensive real-time observations of surface waves is an essential requirement for the advancement of these current studies. However, the balance of requirements for real-time, high-fidelity measurements and large-scale, low-cost measurement device deployment poses a significant challenge. Since the initial recognition of the importance of waves, humanity has been observing them using a variety of means. Within the industry, several mainstream technologies are currently available, including pressure-type wave gauges, acoustic wave meters, accelerometer wave buoys [8], and global navigation satellite system (GNSS) wave buoys. Various types of buoy products have unique advantages compared to other wave measurement devices. However, a significant portion of mainstream buoy products still faces challenges related to deployment difficulties and high costs, which make it challenging to establish large-scale networked observation arrays. For example, the Directional Waverider Mk III produced by the Dutch company Datawell, which has long been regarded as the industry standard, uses a combination of horizontal accelerometers and a compass to measure pitch and roll directly, with the results then being used to calculate the wave direction [9]. Nevertheless, the Directional Waverider Mk III has a diameter that ranges from 0.9 m to 1.1 m and weighs more than 200 kg because of the inclusion of additional batteries for extended working hours and its high-precision measurement units. Therefore, the device often requires the use of shipboard cranes and a team for its deployment and recovery.

To gather comprehensive and real ocean wave data, several countries, including the United States, the United Kingdom, Canada, the Netherlands, and Norway, have been actively developing and deploying wave buoys around their coastal regions to establish a network of monitoring systems. Wave buoys have gained prominence in these systems in recent years and have been incorporated into the Global Drifter Program (GDP) [10]. One of the most suitable observation instruments for performing this task in challenging marine environments is the GNSS wave buoy [11]. The GNSS drifting buoy is equipped with a GNSS receiver that allows it to use the Global Positioning System (GPS). In addition to its capacity to record the geographic coordinates of wave data accurately, this buoy holds significant potential for high-fidelity measurements, global deployment, and low device costs. However, there is still a noticeable absence of mature industrial products that can integrate the advantages of the systems referenced above into one system.

Over the decades since the advent of GNSS satellites, research institutions worldwide have harnessed technology to perform buoy-based wave observations. The prevalent GNSS positioning methods that are currently used on buoys include post-processed kinematic (PPK) technology, real-time kinematic (RTK) technology, precise point positioning (PPP) technology, and Doppler velocity measurement technology [12]. However, these methods come with inherent observational limitations when used for ocean monitoring, including their substantial reliance on satellite and orbital product quality, the challenges involved in establishing and maintaining remote base stations for more distant observations, and inherent observation accuracy limitations. Bender conducted a comparative analysis of the various positioning methods used to measure waves and found that the buoy accuracy of wave parameters measured using the RTK and PPP methods was comparable with that of six-axis accelerometer buoys [13]. However, despite these promising results, GNSS buoys were utilized primarily as mere temporary installations for offshore testing. There was no commercially scalable GNSS buoy product in the industry at that time. As early as 1999, Krogstad’s team developed a buoy based on Doppler velocity measurement principles, but their device still relied on differential GPS signals from nearby ship-based or land-based reference stations [14]. In 2003, the Datawell DWR-G GNSS wave buoy, which also used the Doppler velocity measurement principles, was invented, but it was able to operate independently without the need for a reference base station [15]. The measurement accuracy of this buoy was comparable to traditional accelerometer-based wave buoys, with the added advantage of a high-frequency response. However, the product’s comparatively high cost limits its scalability for global deployment. In 2012, the potential of GPS components applied to drifting buoys was confirmed [16]. Raghukumar conducted validation experiments on the Spotter buoy, demonstrating its potential for wave measurement applications [17]. However, there were issues with high-frequency noise and vertical measurement accuracy in early versions of the Spotter buoy. In 2022, Jim Thomson developed the second version of the microSWIFT buoy, which can be deployed via airdrop. This buoy processes GPS velocity data sampled at 5 Hz to calculate wave spectra. Due to the deployment method, there are still questions about whether the buoy’s design and cost are suitable for large-scale global deployment [18]. In 2015, Jean Rabault’s team demonstrated the feasibility of using IMU modules for wave measurements, marking a new development direction for wave-measuring buoys [19]. Alari developed a novel wave buoy named LainePoiss (LP), which utilizes a microelectromechanical system (MEMS) inertial measurement unit (IMU) to detect surface motion [20]. The LP buoy is designed to be lightweight (weighing only 3.5 kg), making it easy to deploy, and it has ice-resistant capabilities, allowing it to be used in ice-covered waters. However, its maximum measurement frequency is limited to 1.28 Hz, and the use of a gyroscope introduces low-frequency noise issues. In 2022, Tsubasa Kodaira’s team developed and deployed a wave buoy based on MEMS IMU and solar power. However, due to its heavy reliance on solar radiation, its operational lifespan is short, making it unsuitable for long-term observations [21]. According to Rabault, the OpenMetBuoy-v2021, an open-source device, offers a cost-effective solution for wave and drift measurements in sea ice and open ocean. However, the application of the OpenMetBuoy is still limited to sea ice regions [22]. Currently, there is still no reliable, low-cost product available for high-frequency wave measurements.

In this work, we have developed a novel GNSS drifting buoy for surface wave measurements that uses the kinematic extension of the Variometric Approach for Displacement Analysis Stand-alone Engine (Kin-VADASE) velocity measurement method and the first five wave direction spectra method. Unlike traditional approaches, this buoy only requires a GNSS dual-frequency signal receiver, thus eliminating the requirements for additional high-precision measurement units and the development of unique satellite orbital products. This innovative design reduces both power consumption and overall costs significantly while also enabling the deployment of a high-precision measurement capability on a global scale. The remainder of the paper is structured as follows. Section 2 provides insights into the buoy’s hardware and its exterior design, along with technical details with regard to data processing. Section 3 presents the results of two nearshore reliability tests of the buoy, along with comparisons of the data obtained with the corresponding data obtained from mainstream products. Section 4 offers an analysis of the measurement errors, and finally, in Section 5, we summarize our findings, propose areas for future improvement, and outline potential research directions.

2. Technology Description

2.1. Buoy Design

The primary focus of the buoy design is on its ability to measure wave parameters using a GNSS receiver alone. It is of paramount importance that the antenna maintain a position above the water surface to ensure appropriate GNSS signal acquisition. However, a greater hull height may impede the buoy’s ability to follow the motion of short waves because it will have a relatively higher resonance frequency. For example, the Surface Wave Instrument Float with Tracking (SWIFT) buoy, which is designed in a spar shape with a height of up to 1.8 m and a resonance frequency of 0.8 Hz, is unsuitable for high-frequency wave measurement [23]. To strike a balance between these constraints, we opted for an ellipsoidal buoy shape. The buoy’s internal structure, its exterior, and its on-site configuration are depicted in Figure 1 and Figure 2. The novel GNSS buoy has a height of 636 mm and a diameter of 534 mm. The inner section accommodates all the electronic components required, and the outer section provides the necessary buoyancy. The ellipsoidal hull shape leads to excellent wave-following behavior, and the arched roof offers additional space for the inclusion of the GNSS receiver. The roof is polished meticulously and is constructed from water-resistant materials to minimize the likelihood of water accumulation. Additionally, the battery is positioned at the bottom of the buoy, thus serving as ballast. These combined design elements yield a resonance frequency that exceeds 1 Hz and an overall weight of 20 kg.

With regard to the electronic components, we positioned the GNSS receiver at the top of the buoy. We selected the Novatel OEM7500 GNSS receiver (NovAtel, a part of Hexagon, Calgary, AB, Canada) because it can achieve centimeter-level accuracy in land-based RTK applications, which also validates its capability for use in VADASE surface drifting buoy applications. This dual-frequency receiver (L1/L2) can record data simultaneously from GPS, GLONASS, BeiDou [24], and Galileo satellites at frequencies of up to 20 Hz, thus ensuring the acquisition of usable data even under harsh sea conditions. Furthermore, the receiver captures the raw satellite data, including the carrier phase output required to achieve centimeter-level accuracy. Another advantage of using the dual-frequency receiver is to eliminate the most significant measurement errors resulting from the ionospheric delay that occurs during the transmission of the satellite signal. For the current application, the GNSS receiver acquires samples at 5 Hz from the satellite systems mentioned above; this frequency is sufficient for obtaining wave measurements. An Iridium satellite transmission mode is integrated on board that enables data to be transmitted without distance limitations. In future work, the data transmission mode may be transitioned to BeiDou short message communication to reduce the data transmission costs. This mode also supports bidirectional transmission, which allows the sampling period to be adjusted based on commands from the land station. Ordinarily, the buoy is set to sample continuously. Additionally, with the thermal and conductivity sensors that are located at the buoy’s bottom, the device is designed to measure the ocean surface thermohaline features, which will also be explored in future work.

2.2. GNSS Data Analysis

The Variometric Approach for Displacement Analysis Standalone Engine (VADASE) was invented as an innovative GPS data processing approach in recent years [25]. Originally designed for application to seismological data analysis, this method is regarded highly as an effective and cost-efficient analysis solution for obtaining real-time measurements of seismic waves. The VADASE is based on a continuous collection of carrier phase observations from a standalone GPS receiver and on the use of standard GPS broadcast products, including orbits and clocks, in real-time.

However, in seismic wave estimation applications, the emphasis is placed predominantly on low-frequency wave observations. In contrast, in the context of ocean wave measurements, the primary focus is then placed on the observational capability for higher-frequency data, typically at frequencies exceeding 0.1 Hz. In this study, we use a kinematic extension of the variometric approach known as Kin-VADASE [26], which has been adapted specifically for navigation purposes and has been applied previously to wave observations [27]. The characteristics of the Kin-VADASE method mean that it is exceptionally suited to wave measurement applications. The method presented here uses the properties of the VADASE approach to detect 3D displacements over specific time intervals. We acquire the wave velocities when the measured displacements are divided by their corresponding time intervals. When compared with traditional GNSS-based buoy positioning methods, our proposed approach offers higher theoretical positioning accuracy and does not necessitate any additional correction services. A detailed description of the methodology can be found in Appendix A.

2.3. Parameters and Spectral Analysis

We use a sliding window approach that segments the long velocity time series into segments of 20 min each. These time intervals are long enough to encompass not only the high-frequency wave signals but also the low-frequency tide and mesoscale eddy signals. The processing steps for these velocity series are shown in Figure 3. When the corresponding displacement sequences are obtained, we then need to integrate the velocity sequences. During the integration process, the component induced by the low-frequency signals and the system velocity estimation errors is gradually amplified. We can remove this trend directly through linear regression, which results in a three-dimensional displacement sequence that represents the realistic propagation characteristics of the ocean waves closely. The genuine high-frequency motion displacement signals caused by the waves are acquired via band-pass filtering [28]. As demonstrated by the Datawell specification, the DWR-G buoy displacement time series includes a 0.01 Hz high-pass filter to remove low-frequency motions. To make statistical results comparable, a similar filter with the same cut-off frequency is applied to the north, west, and altitude time series of both the Vardase and PPK results. A slight difference in the performance of filters may exist. However, they are conceived to have no significant effect on filtered time series because of the small variance. To eliminate the impact of the buoy’s own resonance frequency on wave observations, we set the threshold of our low-pass filter to 1 Hz. Then, we use a three times standard deviation outlier detection method to remove outliers from the velocity and obtain three-dimensional displacement of waves. This configuration allows us to capture both high-frequency and low-frequency energy distributions within the waves.

The wave direction spectra are computed using the method proposed by Longuet-Higgins and Kuik Ultimately, we derive wave parameters, including the wave height and the wave period from the vertical displacement, east-west displacement, and north-south displacement sequences provided by the method described above. Detailed descriptions of the methods can be found in Appendix B.

3. Effectiveness Tests and Results

To assess the reliability of the proposed buoy, we conducted nearshore buoy performance tests off the coast of China. To facilitate the recovery and debugging of the prototypes, these buoy tests were conducted using a moored setup instead of a drifting approach. Furthermore, after filtering the collected data, the mooring lines between the buoys do not significantly affect the calculation of wave parameters. The test procedure encompassed the observation of buoy-measured wave parameters, including the significant wave height and the mean wave period, along with the wave energy propagation directions. During the test process, we also used the Datawell DWR-G4 (Datawell, Haarlem, The Netherlands), which is regarded as the industry standard. We set up the Datawell DWR-G4 in the same marine environment as our proposed buoy for comparison and deemed the experimental data obtained from the Datawell DWR-G4 to be both controlled and reliable. Initially, we elected to conduct the tests in the South China Sea within the coastal waters of China. Subsequently, to evaluate the buoy’s performance under conditions with less pronounced fluctuations, we conducted further testing in the relatively calm marine environment of Laoshan Bay, Qingdao, China. In all conducted tests, our novel buoy was configured with a sampling frequency of 5 Hz, whereas the Datawell DWR-G4 was set at a sampling frequency of 1.28 Hz.

3.1. South China Sea Test

The test was conducted in the South China Sea at a location approximately 6.6 km offshore from the coast of Guangdong, China, from 28 to 29 October 2020. Our test site is located at the entrance of Bohe Bay (111.25° E, 21.405° N). This location provides an excellent vantage point for observing swells entering the port from the open sea. The geographical coordinates of the testing site, the water depth, and the live setup at the experimental site are illustrated in Figure 4. The test buoy and the Datawell DWR-G4 were moored securely in place within the same area with an approximate separation distance of 10 m.

Because the baseline distance between the buoy and the onshore base station does not exceed 8 km, the estimation error for PPK positioning remains below 0.02 m. Consequently, the position obtained via the PPK approach can be considered to be the true value of the buoy’s position. We selected a 5 min segment of real vertical and horizontal displacement sequences from this test to perform a comparison between the Kin-VADASE and PPK methods, with results as illustrated in Figure 5. The top panel represents the east-west displacements, and the bottom panel represents the vertical displacements. The horizontal displacements show excellent agreement between the VADASE and PPK results, with a root mean square error (RMSE) of 0.01 m. Although the consistency in the vertical displacement results is slightly inferior, it is still highly satisfactory, with an RMSE of 0.03 m. This relatively minor inconsistency can be attributed to the inherently less accurate nature of measuring vertical displacement through GNSS technology, a characteristic shared by both the PPK and VADASE methods. Nevertheless, the high level of consistency overall underscores the precision of the VADASE method here and indicates its potential applicability to ocean wave monitoring.

The spectral wave statistics obtained from both the Datawell buoy and the GNSS buoy when using the two methods for a 30 min data recording period are compared in Figure 6. The wave spectra show two prominent peaks, which indicate that each wave field consists of a swell that is transported into the port and a locally generated wind wave. The spectra generated using the PPK and VADASE methods demonstrate remarkable consistency across the majority of the wave frequency band. Although there is a notable low-frequency discrepancy between the VADASE and PPK results, this discrepancy is inconsequential during the wave parameter calculation process and does not affect the statistics significantly. All three spectral results identify the two frequency peaks successfully, thus indicating that they are capable of measuring both the swell and the wind waves. Both the VADASE and PPK methods cover the entire frequency band and display clear frequency shapes beyond the wave frequency range of up to 1 Hz. However, the Datawell spectrum was cut off at 0.64 Hz because of its relatively low sampling frequency of 1.28 Hz. The GNSS buoy results show that the wind wave band extends up to approximately 0.80 Hz, thus indicating that the Datawell buoy may miss a proportion of the high-frequency range of the wind waves. At frequencies in excess of 1.00 Hz, peaks are shown in both the VADASE and PPK spectra that are likely to be attributable to buoy resonance oscillations and are filtered out in the subsequent analysis. Upon closer examination of the spectra, the peak frequency obtained from the VADASE method is slightly lower than that obtained from the PPK method. This discrepancy can be partially attributed to sampling inconsistencies between the buoy when using the VADASE and Datawell methods. Because of the differences in their sampling frequencies, the frequency ranges of these spectra differ. Although they experience the same peak frequency, their mean periods will differ due to the different frequency ranges they observed.

Comparisons of the bulk wave parameter estimates from the VADASE and PPP methods are presented in Figure 7. Scatter diagrams are used to depict estimates of the significant wave height (Hs), the mean wave period (Tm), and the mean wave direction (D) obtained from the VADASE method and compare them with corresponding estimates from PPK results from the same buoy. These fundamental wave parameters are largely influenced by the wave spectra, and this leads to excellent consistency between the estimates of Hs, Tm, and D from the VADASE and PPK methods. This indicates that the VADASE method is capable of measuring waves and is, in a statistical sense at least, as accurate as the PPK method, with correlation coefficients that all exceed 0.95 and RMSE values of 0.03 m, 0.12 s, and 1.05°, respectively. The estimates of Hs, Tm, and D obtained from the buoy using VADASE and the Datawell buoy are compared in the scatter diagrams in Figure 8. The comparison of the mean period and significant wave height time series between VADASE and the Datawell buoy is shown in Figure 9. The Hs results show strong consistency between the VADASE and Datawell buoys, with a correlation coefficient of 0.91 and an RMSE of 0.05 m. The maximum bias of 0.12 m is acceptable for wave height measurements. In addition, D also exhibits a reasonable correlation between the methods, with a coefficient of 0.68. However, because the waves during the field test mainly originate from the port mouth, their directions do not vary significantly. Minor deviations between the two buoys can result from slight interruptions, which ultimately affect D. Under these conditions, it is then more meaningful to consider the root mean square error (RMSE), which is 2.49°; this suggests that both buoys capture the peak direction. The largest deviation is approximately 10°, which is partially due to the limited resolution of directional measurement. The Tm correlation between the two buoys is 0.67, which is comparable to the results of previous research findings. However, there is a bias of 0.7 s and an RMSE of 0.75 s, which indicates that the buoy using VADASE underestimates Tm when compared with the Datawell buoy. This discrepancy warrants further analysis and will be addressed later.

3.2. Laoshan Bay Test

Laoshan Bay is a naturally formed bay that is sheltered by the surrounding topography. These physical characteristics reduce the direct impact of external wind waves. Together with its stable weather patterns, these properties make Laoshan Bay an ideal testing site. During the period from 7 to 10 November 2022, we conducted a buoy test array deployment near an offshore platform located within Laoshan Bay. In this test, we tested a total of eight buoys divided into two groups of four. This deployment strategy involved connecting the test buoys into a linear configuration using floating ropes with a 5 m separation distance between adjacent buoys. Floating balls were then attached between the buoys and anchors to minimize the rope tension acting on the buoys. The distance between the anchor and the buoy on the array head (where the anchor is, of course, at the end) is more than 1.5 times the depth of the measurement area. In addition, we also deployed a Datawell DWR-G4 buoy for comparison purposes near the measurement array. Under the relatively calm wind and wave conditions, the ropes between the buoys would not become entangled. It should also be emphasized that the deployment of the equipment for this test required only a small motorboat and a two-person team, resulting in a significant reduction in deployment costs when compared with conventional equipment deployment. The geographical coordinates of the test site, the water depth, and an illustration of the live experimental site are shown in Figure 10.

We selected two representative test buoys (designated Testbuoy1 and Testbuoy2) and compared their measurement data with those from the Datawell buoy in terms of the significant wave height (Hs), the mean wave period (Tm), and the dominant wave direction (D). The Hs results are presented in Figure 11, the Tm01 results are presented in Figure 12, and the D results are presented in Figure 13. We can clearly see that even in the presence of mild wind and wave conditions, the testing buoys and the Datawell buoy show strong consistency in both their significant wave height and main wave direction characteristics, particularly during processes that involve changes in wave energy direction. However, a notable difference emerges in the wave mean period statistics. Closer examination reveals that the Datawell buoy cannot resolve waves with frequencies higher than 0.64 Hz because of its relatively low-frequency sampling rate. In contrast, the VADASE buoy, with its high-frequency sampling rate and appropriate hull design, can record wave motions accurately up to 1.00 Hz. The remaining differences are presumed to be caused by variations in the hull responses and wave inhomogeneity and will require further investigation. This result also underscores the significance of the hull design and the data processing algorithms, in addition to sensor positioning accuracy, in the determination of the buoy performance during contemporary wave monitoring.

4. Discussion

During the test process, we observed a disparity between the mean wave period values estimated by the novel buoy and the Datawell buoy. This discrepancy may partially be caused by the difference between the sampling frequencies of the two buoys. This signal loss is manifested in the power spectra calculations. Although they are effective in extending the buoy working times, lower sampling frequencies tend to lose high-frequency information. According to the formulas for the wave parameters, the wave height computation depends solely on the zero-order moment, whereas the period calculation necessitates the division of the zero-order moment by the first-order moment. The process of calculation of the wave parameters amplifies the limitation caused by the absence of high-frequency signals. Therefore, the disparate sampling frequencies used by the two devices contributed part of the observed difference between the mean wave periods of the Datawell buoy and our test buoy.

To validate this assertion, we require a controllable input signal that can simulate different sea conditions. In this work, we used the Jonswap waves [29] as input signals to provide better simulations of complex ocean waves under different wind speed conditions. We generated a set of input signals with wind speeds that ranged from 0 to 30 m/s at intervals of 0.5 m/s, giving a total of 60 gradient-distributed wind speeds. The waves under each wind speed condition are sampled at rates of 1 Hz, 1.28 Hz (corresponding to the Datawell DWR-G4 sampling rate), 2.5 Hz, 3.5 Hz, and 5 Hz, thus allowing us to investigate the influence of the sampling frequency on the mean wave period calculations under conditions of identical input signals and gradient-distributed wind speeds. Power spectra comparisons of the signals at the different sampling frequencies are shown in Figure 14, and comparisons of the mean period variations of these spectra at the different wind speeds are depicted in Figure 15. Notably, as illustrated in Figure 15b, there are significant differences between the results obtained at the lower sampling frequencies, e.g., 1 Hz and 1.28 Hz, and those at the higher signal frequencies; in addition, these differences increase gradually with increasing wind speed. This illustrates that as the wind speed increases, the influence of the high-frequency signals becomes more pronounced. The average difference in Tm between the sampling frequencies of 1.28 Hz and 5 Hz is approximately 0.28 s. This result underscores the critical influence of the high-frequency energy at frequencies above 1 Hz on the mean wave period calculations. In addition, it indirectly affirms that it is necessary for the wave measurement buoys to operate in a high-frequency sampling mode. The work performed here shows that the calculation process for the mean period is sensitive to the higher frequency signals and that acquisition of these high-frequency signals is necessary to achieve results that are close to the real wave situation.

5. Conclusions

GNSS positioning modules have found widespread applications across various offshore platforms worldwide, including buoys, marine vessels, and large offshore installations, for purposes that include attitude correction and wave data collection. However, the wider application of GNSS technology in the maritime field remains to be developed for drifting buoys by many countries. This paper introduces a new GNSS drifting buoy design that stands out from traditional buoys, such as accelerometer buoys, by eliminating the requirement for additional high-precision measurement units. This not only reduces the overall cost but also minimizes the volume and weight of the buoy, which is advantageous for global deployment. These attributes mean that the buoy is well-suited to construction within an oceanic network and enables large-scale network observation of wave parameters. In terms of the calculation module, we used the Kin-VADASE method coupled with a matching algorithm for the oceanic elements and the directional spectra. The combination of this approach with high-frequency data acquisition enables a more accurate representation of the ocean waves. The newly designed buoy has an optimized internal structure that enables it to accommodate larger battery capacities and thus prolongs its operational lifetime. Additionally, the buoy’s size and weight have been controlled to enable its placement by just one or two individuals, thus reducing the deployment cost.

With regard to the measurement precision, we initially tested the accuracy of the Kin-VADASE method. By comparison with the precise point positioning (PPP) method, we confirmed that the Kin-VADASE method can deliver high-quality results for the wave element and directional spectra calculations, even after the demands placed on satellite data quality and quantity were reduced. To perform the in situ tests, we selected two coastal regions, comprising the South China Sea and the Laoshan Bay area. The test results were compared with those from the industry-standard product, i.e., the Datawell DWR-G4 buoy, and showed favorable agreement in terms of the significant wave height, the mean wave period, and the main wave direction. After multiple tests, the feasibility of our developed GNSS wave measurement method with the novel buoy was confirmed. Comparison with internationally recognized wave measurement buoy products showed that the buoy’s technical specifications for parameters, including the significant wave height (Hs) and the dominant wave direction (D), are on a par with those of the leading international products, thus demonstrating significant potential for global deployment of the proposed buoy. However, there is an obvious disparity between the estimated mean wave period (Tm) values from the two sets of data acquired from the novel buoy and the Datawell buoy, which is caused by the sampling frequency difference between the two products. In the error analysis experiment, the disparity in estimation of the mean wave period was shown to become more pronounced with increasing wind speed. This result explains more than a third of the discrepancy between the Tm values acquired in offshore tests. At present, the buoy product design is being improved continually in areas including the power supply, algorithm refinement, and cost control. We are committed to conducting further offshore tests and reliability assessments under harsh sea conditions, which will require a diverse range of experimental conditions and will be pursued in future work.