Localizing unknown nodes with an FPGA-enhanced edge computing UAV in wireless sensor networks: Implementation and evaluation

Pervasive and Mobile Computing 103 (2024) 101961 Available online 19 June 2024 1574-1192/© 2024 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Contents lists available at ScienceDirect Pervasive and Mobile Computing journal homepage: www.elsevier.com/locate/pmc Localizing unknown nodes with an FPGA-enhanced edge computing UAV in wireless sensor networks: Implementation and evaluation Rahma Mani a,b,∗, Antonio Rios-Navarro b, Jose Luis Sevillano Ramos b, Noureddine Liouane a aResearch Laboratory of Automatic Signal and Image Processing (LARATSI), National Engineering School of Monastir, University of Monastir, Monastir, 5019, Tunisia bDepartment of Computer Technology and Architecture, University of Seville, Seville, 41012, Spain ARTICLE INFO Keywords: Field Programmable Gate Array (FPGA) Localization DV-Hop Kalman filter Bounding box Mobile Sensor Networks Edge Robotics UAVs ABSTRACT Great interest is directed toward real-time applications to determine the exact location of sensor nodes deployed in an area of interest. In this paper, we present a novel approach using a combination of the Kalman filter and regularized bounding box method for localizing unknown nodes in an area using an FPGA-enhanced edge computing UAV whose trajectory is known and is represented as the position of many anchors. The UAV is equipped with a GPS system that allows it to gather location data of sensor nodes as it moves around its environment. We employ a regularized bounding box to predict the positions of the unknown nodes using regularization factors and we use the Kalman filter algorithm to smooth and improve the accuracy of the sensor nodes to be localized. In order to localize the unknown nodes, the UAV receives the number of hops from each node and uses this information as input to the localization algorithm. Furthermore, the use of an FPGA board allows for real-time processing of sensory data, enabling the UAV to make fast and accurate decisions in dynamic environments. The localization algorithm was implemented on the FPGA board ‘‘Zynq MiniZed 7007s evaluation board’’ using Xilinx blocks in Simulink, and the generated code was converted into VHDL using Xilinx System Generator. The algorithm was simulated and synthesized using ‘‘Vivado’’ software. In fact, the proposed system was evaluated by comparing the performances achieved through two different implementations: Hardware and Software implementation. In effect, the performance of FPGA hardware implementation presents a new achievement in localization due to its easy testing and fast implementation. Our results show that this approach can efficiently locate unknown nodes with good latency and high accuracy. In fact, the execution time of the FPGA-integrated algorithm is reduced by about 60 times compared to the software implementation and the power consumption is about 100 mW, which proves the suitability of FPGA for localization in WSNs, offering a promising solution for various mobile WSN applications. 1. Introduction In the emergent wireless sensor network, accurate localization of sensor nodes is a crucial challenge for many potential applications, such as smart cities, intelligent navigation, process supervision and control, Internet of Things (IoT) devices and military tracking applications [1–3]. For example, in military application Fig. 1, it is often necessary to determine the position of a dozen ∗Corresponding author. E-mail address: [email protected] (R. Mani). https://doi.org/10.1016/j.pmcj.2024.101961 Received 14 September 2023; Received in revised form 13 June 2024; Accepted 15 June 2024 Pervasive and Mobile Computing 103 (2024) 101961 2 R. Mani et al. Fig. 1. Example of mobile UAV for WSN localization in military application. or even hundreds of soldiers in a war zone for rescue operations, where each soldier is equipped with wireless LORA or Zigbee technology sensors to communicate with a drone in order to locate the soldiers or gather information about the environment [4]. In the literature, a limited number of research works have been carried out to implement localization algorithms on a dedicated hardware architecture (see Related Works section). In this work, we therefore present a complete procedure for implementing and evaluating a Regularized Bounding-Box within the Kalman Filter algorithm (RBBLA-KF) for localizing unknown nodes in an area using an FPGA-enhanced mobile UAV that employs edge computing to perform real-time processing and analysis of sensory data without relying on remote servers or cloud-based computing resources, which brings computational tasks and data storage closer to the sources of data (Fig. 2), unlike traditional cloud computing, where data is collected and processed centrally. This can reduce latency since data does not have to be transmitted over long distances to a remote server [5–7]. Also, it can improve performance, since data can be processed and analyzed closer to the source. Moreover, it can increase the reliability and security of the system since data is not transmitted over a network and is, therefore, less vulnerable to interference or attacks [8,9]. Using an FPGA board allows for real-time processing of sensory data, enabling the UAV to make fast and accurate decisions in dynamic environments. This makes the approach particularly well-suited for a wide range of applications that require low latency, high reliability and accuracy, such as real-time control systems, remote health monitoring, and industrial automation. It is also well-suited to environments with limited or unreliable connectivity, such as remote or inaccessible areas. In our case, the main task of edge computing is resolving the localization problem in large-scale WSNs. The proposed localization algorithm uses a combination of the Kalman filter and the regularized bounding box method to accurately determine the positions of unknown nodes within an environment. The Kalman filter is a widely used method for estimating the state of a system from noisy, incomplete, or uncertain measurements [10], while the regularized bounding box method uses a nonlinear method to determine the location of wireless sensor nodes. This is an essential capability for an UAV operating in an irregular environment, as it allows the UAV to navigate and perform tasks without depending on pre-mapped information. The algorithm was designed and simulated in Matlab/Simulink, then we translated it to VHDL for hardware implementation via Xilinx System Generator (XSG) tools. The simulation, synthesis, and implementation are carried out using Vivado software. After that, we implemented the design and got the bitstream file to program the FPGA board. The physical system illustrated in Fig. 3 shows the Zynq FPGA-SoC diagram in which the ARM core is the AXI master and the FPGA is the AXI slave, where the AXI Lite protocol communicates between the PS module and the PL module. Since we used a Zynq platform to deploy our algorithm in the PL (Programmable Logic), we designed a C software code to interface the data collected from the WSN and our algorithm. The C code runs in the PS (Processing System) block of the Zynq architecture. We performed the execution in real-time on the board to test the algorithm’s performance and accuracy. Our results showed that the algorithm was able to effectively localize unknown nodes, with a high degree of accuracy. Also, this approach can effectively and efficiently localize unknown nodes within the UAV’s environment, providing a promising solution for numerous applications. Overall, in this paper, the motivation behind this research is to address the challenges posed by dedicated localization systems for WSNs with mobile anchor and the implementation of localization algorithm on an FPGA board, as well as improving localization accuracy in the case of range-free WSNs. Our contribution consists of examining the constraints and suggesting improvements for using the combination of the regularized bounding-box algorithm with the Kalman filter to improve the average localization error. In addition, we propose the practical implementation of the algorithm proposed in this study on a reconfigurable FPGA architecture aimed at minimizing energy consumption and accelerating computation using the Matlab-XSG and Vivado-HLS software tools. The rest of the paper is organized as follows: After the introduction in Section 1, Section 2provides a review of related works in the field. Section 3describes the regularized Bounding Box boosted Kalman filter algorithm for localization in WSN. Section 4presents the results of the simulations for RBBLA-KF algorithm and the qualitative evaluation of the performances using representative statistics. Section 5summarizes the implementation and design of the proposed system on Matlab-XSG and FPGA, with a focus on the comparative study between fixed-point and floating-point implementations. Finally, we discuss and conclude our work in Section 6. Pervasive and Mobile Computing 103 (2024) 101961 3 R. Mani et al. Fig. 2. Edge-computing architecture. Fig. 3. Zynq physical system block diagram. 2. Related works There are several existing works that are related to localization algorithms [11–13] also the implementation of different algorithms on FPGA boards [14,15] and the exploitation of edge computing [16–18] with mobile robots [19,20] and UAVs [21,22]. Previous efforts have focused on developing localization algorithms [23], employing techniques such as the Kalman filter [24], visual tracking, and sensor fusion to determine the robot’s position and the positions of surrounding sensor nodes [25]. Paper [21] presents localization techniques via mobile anchors dedicated to WSNs, eliminating all human intervention and configuration of the deployment zone. Indeed, the basic theoretical and algorithmic concepts of UAV localization are exhaustively presented and structured. Also, the issues addressed and discussed throughout this paper are mobility models, computational complexity and localization performances. In [22], the authors address the problem of jointly programming communications and mobility speed control in assisted wireless sensor networks. Indeed, a deep reinforcement learning approach jointly optimizes UAV speed and data reception quality while minimizing data loss rate and energy consumption. In addition, a scheduling algorithm based on multi- UAV deep reinforcement learning is proposed. It is clear from the above literature that researchers are continually improving and optimizing localization methods through research and development activities. However, most of them focus on planar localization. The challenging three-dimensional spatial environment is more demanding in terms of positioning algorithms, spatial complexity Pervasive and Mobile Computing 103 (2024) 101961 4 R. Mani et al. Fig. 4. Sensor node anatomy equipped by a SoC-FPGA. and implementation requirements than the relatively simple two-dimensional planar positioning environment. The improvement and optimization of our algorithm and its implementation in 3D achieve the desired objective for WSN applications. Additionally, there have been trials to implement mobile robot algorithms on FPGA boards [26–28], offering advantages such as high-speed processing, low power consumption, and easy modification and updating of the algorithms. Paper [29] discusses a centralized localization method in which a mobile anchor estimates node positions based on the received signal strength indicator (RSSI) and the angle of arrival (AoA) information. In fact, a mobile anchor equipped with a reconfigurable FPGA architecture moves along a random path in the network and performs a clustering of certain neighboring nodes with an RSSI level exceeding a prescribed threshold using a smart antenna. The coprocessor (FPGA) estimates location information (distance and angle) for unknown nodes at two distinct reference points and stores it in a table. The FPGA then calculates the positions of nodes encountered twice in the array. The performance of the presented system is tested on a dedicated FPGA board with several fixed-point hardware simulations to assess localization accuracy, computational complexities and energy consumption. However, in this paper [29], the authors treat the localization problem in the 2D deployment zone, which does not reflect the reality of the 3D WSN applications. Paper [30] presents the combination of the 3D DV-HOP positioning algorithm with multiple communication radii (MCR) and then the Particle Swarm Optimization (PSO) algorithm was used to obtain the optimized Hop distance correction. Their work is considered algorithmically complicated because it involves PSO and an inversive matrix, which makes it difficult and challenging to implement on FPGAs. The paper [31] deals with the complexity of implementing advanced algorithms on programmable architectures of the FPGA chip dedicated to mobile robots. The authors have explored the various applications in which FPGAs have been successfully integrated and have discussed their operating advantages and limitations, including improvements in real-time processing, energy efficiency and adaptability to environments In the context of WSNs, the utilization of FPGAs has emerged as a crucial and beneficial approach. WSNs typically rely on microcontrollers as the processing core to handle application tasks. However, microcontrollers often have limitations in terms of computational effort, memory size and energy consumption. This is where FPGAs come into play, offering a range of advantages that make them suitable for WSNs. Given the limited capacity of the microcontroller (CPU) and memory resources built into a sensor node, the FPGA is used as a coprocessor to help the CPU execute certain tasks requiring high algorithmic complexity, such as the localization algorithm, while speeding up the calculation process. Fig. 4 shows the anatomy of the sensor node equipped with an FPGA core. The sensor gathers informations about the environment and converts the signals via the ADC converter before transmitting them to the microcontroller. The wireless transceiver chip, radio module and communication protocol are required to exchange informations with neighboring sensor nodes. The microcontroller processes the sensor signals, determines the appropriate minimum number of hop-counts, communicates the number of hops to the FPGA module via the AXI Lite protocol, and manages local memory and I/O data. The FPGA module calculates the location of the sensor node via the our algorithm. Finally, the power supply unit manages and optimizes energy consumption from the battery. One key advantage of FPGAs in WSNs is their reconfigurability, allowing them to adapt to changing requirements and tasks. Considering the aforementioned benefits, it becomes evident that FPGAs have become an appropriate component in the design and implementation of WSNs. They empower WSNs to overcome the limitations of microcontrollers and efficiently handle complex computational tasks. By leveraging the re-programmability, real-time processing, design reusability, and hardware acceleration of FPGAs, WSNs can achieve enhanced functionality, improved performance, and extended operational lifetimes. Building upon these endeavours, our contribution extends these approaches by implementing a localization algorithm on an FPGA board for an edgecomputing UAV in 3D WSN applications. This empowers the UAV to perform localization tasks in real-time, utilizing local resources and saving energy. Pervasive and Mobile Computing 103 (2024) 101961 5 R. Mani et al. Fig. 5. Recursive Bounding-Box algorithm. 3. The regularized BBLA-KF algorithm for localization in WSN Based on [24], which uses the mobile bounding box algorithm for localization problems in 3D WSN, the localization solution would be further improved by using regularization factors for the bounding box algorithm, which we then refined with the Kalman filter (RBBLA-KF) to accurately determine the positions of unknown nodes in an area. In Fig. 5, at the anchor position 𝐴𝑘,𝐵𝐵𝑘 denotes the bounding box of an unknown node position where 𝐵𝐵𝑘−1 denotes the resulting bounding box delimiting the unknown node for the old positions of the mobile anchor and 𝐴𝐵𝑘denotes current bounding box. The iterative rules for updating the bounding box are described by the following equation: 𝐵𝐵𝑘=𝐴𝐵1∩𝐴𝐵2...𝐴𝐵𝑘−1 ∩𝐴𝐵𝑘,∀𝑘= 1..𝑁𝑎 =𝐵𝐵𝑘−1 ∩𝐴𝐵𝑘(1) then, 𝐵𝐵𝑘=⎡ ⎢ ⎢ ⎢ ⎣ 𝑥− 𝑘𝑥+ 𝑘 𝑦− 𝑘𝑦+ 𝑘 𝑧− 𝑘𝑧+ 𝑘 ⎤ ⎥ ⎥ ⎥ ⎦ (2) Where 𝑁𝑎is the total number of the mobile anchor positions. In order to improve the localization algorithm, we introduce regularization factors that further localize unknown nodes and minimize the boundary node error for each coordinate (𝑥, 𝑦, 𝑧)of the sensor node. The location of the unknown node at the anchor position 𝑘is calculated via the weighted centroid algorithm given by the next equation: 𝑋𝑚𝑒𝑠 =⎡ ⎢ ⎢ ⎢ ⎣ 𝛼𝑥1 − 𝛼𝑥0 0 0 0 0 0 𝛼𝑦1 − 𝛼𝑦0 0 0 0 0 0 𝛼𝑧1 − 𝛼𝑧 ⎤ ⎥ ⎥ ⎥ ⎦ ⋅ ⎡ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎣ 𝑥+ 𝑥− 𝑦+ 𝑦− 𝑧+ 𝑧− ⎤ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎦ (3) where 𝛼𝑥, 𝛼𝑦, 𝛼𝑧represent the regularization factors to be adjusted and depend on the regularity of the deployment zone (isotropic or anisotropic case) and the trajectory of the mobile anchor. In our simulation, these parameters are set by a trial-and-error procedure. The bounding box is defined by a set of coordinates that enclose an object and is typically represented as a set of six geographical coordinates for their three-dimensional field: the 𝑥−is the left side, 𝑥+is the right side, 𝑦−is the bottom side, 𝑦+is the top side, 𝑧− is the back side and 𝑧+is the front side of the bounded box, which are then processed to determine the unknown nodes’ positions. One of the main advantages of using bounding boxes for object localization is that they are simple and efficient to compute [32,33]. The bounding box is also highly customizable, quickly and easily generated and can be used in a wide range of applications. In addition to the regularized bounding box, we also used the commonly used method Kalman filter, for localizing and identifying sensor nodes. The Kalman filter uses a mathematical model of the system to combine the available measurements with our prior knowledge of the system’s state, given by the next equation, to produce an estimate of the current state. 𝑋𝑘=𝐴𝑋𝑘−1 +𝑤𝑘(4) Pervasive and Mobile Computing 103 (2024) 101961 6 R. Mani et al. Fig. 6. The proposed RBBLA-KF algorithm to be implemented on FPGA. wherein 𝑋𝑘= (𝑥𝑘, 𝑦𝑘, 𝑧𝑘)𝑇is the estimated position of the unknown node when the anchor is at station 𝑘, the vector of process noise 𝑤𝑘considers the perturbations on the position of the unknown node. In this scenario, we consider the position of the unknown node as fixed and unaltered during the localization process, so the value of 𝑤𝑘is null. The transition matrix 𝐴3×3 = Identity Matrix. For this system, the measurement equation is visualized as follows: 𝑍𝑚𝑒𝑠 𝑘=𝐻𝑋𝑚𝑒𝑠 𝑘+𝑣𝑘(5) 𝑣𝑘 ∼𝑁(0, 𝑅𝑘) Where 𝑋𝑚𝑒𝑠 𝑘= (𝑥𝑚𝑒𝑠, 𝑦𝑚𝑒𝑠, 𝑧𝑚𝑒𝑠)𝑇is the measurement vector given by the moving bounding box process. 𝐻is the observation matrix with 𝐻3×3= Identity Matrix, and 𝑣𝑘is the observation noise vector assumed to be a vector of normally distributed random variables with zero mean and covariance matrix 𝑅. In the case of UAV, the Kalman filter uses the UAV’s GPS measurements and the number of hops received from each unknown node to estimate the positions of the unknown nodes. The GPS measurements provide information about the UAV’s own position, while the number of hops provides information about the distance between the UAV and the unknown nodes. By combining these two types of information, our algorithm is able to produce accurate estimates of the unknown nodes’ positions. They can also be tailored to the specific needs of a given application. Together, the Kalman filter and regularized bounding box method provide a robust and accurate approach for localization, allowing the UAV to navigate and perform tasks in unknown environments. The Kalman filter provides reliable estimates of the unknown nodes’ positions, while the regularized bounding box method provides the bounding box limits for each unknown node. This combination of techniques allows the UAV to accurately locate the surrounding nodes, enabling it to navigate and perform tasks with confidence. Fig. 6 gives the algorithm structure of the proposed RBBLA-KF. The box on the left first describes the flooding phase, which consists of finding the number of hops separating the mobile anchor and the unknown nodes. Next, the regularized mobile bounding box algorithm is applied via the Eqs. (1) (2) (3). The box on the right presents the Kalman filter algorithm for the refinement process to improve the estimated position given by the regularized mobile bounding box. 4. Simulation results The performance of our RBBLA-KF algorithm is being evaluated through studies on the properties of a 3D wireless sensor network. The average localization error (ALE) is presented to analyze the impact of varying the regularization factors on the results of sensor node localization. 𝐴𝐿𝐸 =∑𝑁 1√(𝑥𝑘,𝑚𝑒𝑠 −𝑥𝑘)2+ (𝑦𝑘,𝑚𝑒𝑠 −𝑦𝑘)2+ (𝑧𝑘,𝑚𝑒𝑠 −𝑧𝑘)2 𝑅×𝑁(6) 4.1. Regularization factors effects In order to improve the accuracy of sensor node localization in our case study, we studied the effects of varying regularization factors (𝛼𝑥,𝛼𝑦,𝛼𝑧) on the average localization error. The results presented in this section compare the performance of the proposed method under the same conditions of the 3D WSN. We will conduct experiments with 400 sensor nodes randomly distributed across a 100 m ×100 m ×100 m detection zone without obstacles. In order to ensure the drone’s connectivity with the WSN, we set the communication range at 30 m, respecting the condition given by Eq. (7). 𝑅≥𝑍𝑈𝐴𝑉 𝑚𝑎𝑥 (7) Pervasive and Mobile Computing 103 (2024) 101961 7 R. Mani et al. Table 1 Simulation parameters. Parameter Value Deployment area 100 m×100 m×100 m Communication range 30 m Node amount 400 UAV trajectory RWP with 𝑍𝑈𝐴𝑉 𝑚𝑎𝑥 ≤30 m Table 2 𝛼𝑧variation (𝛼𝑦= 0.5;𝛼𝑥= 0.5). 𝛼𝑧0 0.25 0.5 0.75 1 ALE 1.051 1.780 0.6209 0.3066 0.1889 Table 3 𝛼𝑦variation (𝛼𝑧= 1;𝛼𝑥= 0.5). 𝛼𝑦0 0.25 0.5 0.75 1 ALE 0.3869 0.3064 0.1949 0.1006 0.1358 Table 4 𝛼𝑥variation (𝛼𝑧= 1;𝛼𝑦= 0.75). 𝛼𝑥0 0.25 0.5 0.75 1 ALE 0.2376 0.1570 0.1111 0.0979 0.1532 where 𝑍𝑈𝐴𝑉 𝑚𝑎𝑥 represents the height of the drone. Table 1 shows the parameters of the 3D wireless sensor network being simulated, such as deployment area, communication range, number of unknown nodes and UAV trajectory planning. We used MATLAB-2016 software to determine performance measurement simulations. The path-planning of the UAV is a random waypoint trajectory (RWP). We performed various simulations, as mentioned in Tables 2,3, and 4, to determine the best balance between accuracy and RBBLA-KF regularization factors. The optimal regularization factor for the 𝑧-axis equal to one is explainable by the unique spatial constraints imposed by the UAV’s positioning relative to the sensors. In this context, the regularization along the 𝑧-axis eliminates the upper part of the bounding box. Given that the UAV is situated above the sensors during tracking, the upper side of the bounding box is inherently limited by the presence of the UAV itself. Consequently, setting the regularization factor to 1 effectively eliminates the need for an upper limitation in this dimension, aligning the bounding box more precisely with the physical constraints of the UAV’s environment and optimizing the localization algorithm for improved performance. The sample results illustrated in Fig. 7,8 and 9were obtained for the ideal combination of regularization factors (𝛼𝑧= 1,𝛼𝑦= 0.75,𝛼𝑥= 0.75) for evaluating the RBBLA-KF algorithm in different scenarios. The algorithm is evaluated in various deployment zone types: 3D Cubic field, 3D C-shaped field, 3D A-shaped field. These fields are chosen as a combination of isotropic and anisotropic 3D areas. In these figures, the green lines represent the UAV’s trajectory, and the green dots represent the UAV’s positions during the localization process. The blue circles ‘o’ represent the estimated positions of the unknown nodes given by RBBLA-KF and the orange stars ’*’ represent the real positions of the sensor nodes. The red interrupted lines represent the estimated errors. 4.2. Qualitative performances In this section, in order to assess the performance of the RBBLA-KF algorithm versus the BBLA-KF algorithm, as well as to understand the reliability of the statistical results obtained and the robustness of the algorithm, fifty-one independent runs were carried out for various WSN configurations. Fig. 10 depicts the boxplots simulation results that show the impact of the number of nodes with a communication range equal to 30 m and the regularization factors fixed to (𝛼𝑥= 0.75,𝛼𝑦= 0.75,𝛼𝑧= 1). The ALE for RBBLA-KF mentioned in Fig. 10(b) is inferior to that of BBLA-KF illustrated in Fig. 10(a) for the same total number of nodes which confirms that the impact of the regularization procedure can improve localization errors with different scenarios regarding the number of sensor nodes. Fig. 11 shows the empirical cumulative distribution function (ECDF) of ALE. We note that the performance of RBBLA-KF surpasses the results of BBLA-KF. Indeed, Fig. 11(b) gives the cumulative probability for RBBLA-KF’s ALE where 90% of unknown sensor nodes are located with an average localization error lower than 0.26×R meters, whereas Fig. 11(a) for BBLA-KF’s cumulative probability gives an accuracy equal to 0.29×R meters for the same percentage of unknown nodes to be localized. We also find that as the number of nodes increase, the ECDFs tend to decrease. This may be due to the growth in network connectivity, which is intended to give an accurate estimate of the distance between nodes in large-scale WSNs. In conclusion, we can say that increasing the density of nodes improves the performance of the algorithm by improving the accuracy. Table 5 shows the statistical results in the form of mean, standard deviation (Std), maximum and minimum of the ALE results for fifty-one independent simulation runs. In fact, this table proves the superiority of RBBLA-KF over BBLA-KF. Pervasive and Mobile Computing 103 (2024) 101961 8 R. Mani et al. Fig. 7. Localization results of our proposed algorithm in 3D Cubic field. Fig. 8. Localization results of our proposed algorithm in 3D C-shaped field. Fig. 9. Localization results of our proposed algorithm in 3D A-shaped field. Pervasive and Mobile Computing 103 (2024) 101961 9 R. Mani et al. Fig. 10. Average localization error versus the total number of sensor nodes. Fig. 11. ECDF ALE for various number of sensor nodes. Table 5 Statistical results for different number of nodes. Number of nodes 100 200 300 400 500 600 700 800 BBLA-KF Mean 0.2657 0.2596 0.2600 0.2568 0.2613 0.2565 0.2554 0.2572 Std 0.0247 0.0172 0.0178 0.0202 0.0201 0.0170 0.0155 0.0171 Min 0.2208 0.2263 0.2264 0.2195 0.2230 0.2258 0.2270 0.2242 Max 0.3494 0.3097 0.3139 0.3176 0.3193 0.3016 0.2879 0.2975 RBBLA-KF Mean 0.2309 0.2280 0.2290 0.2268 0.2298 0.2207 0.2277 0.2246 Std 0.0295 0.0269 0.0223 0.0169 0.0188 0.0171 0.0196 0.0135 Min 0.1826 0.1822 0.1937 0.1938 0.1973 0.1907 0.1884 0.1962 Max 0.3033 0.3001 0.3039 0.2756 0.2868 0.2706 0.3049 0.2654 Afterward, we examine the impact of varying communication range on the average location error (ALE) for 400 sensor nodes deployed in the 3D WSN. The communication range of sensor nodes extends from 20 to 40 m. The boxplots given in Fig. 12 present the effect of the communication range variation on the ALE. The comparison shows that the ALE of RBBLA-KF (Fig. 12(b)) is lower than BBLA-KF (Fig. 12(a)) for each value in the communication range. It can be clearly seen that as the communication range Pervasive and Mobile Computing 103 (2024) 101961 16 R. Mani et al. Fig. 21. Real-time implementation of localization algorithm. approach on the SoC-FPGA Minized evaluation kit using the Xilinx system generator and Vivado HLS software was successful. The performance of the system was evaluated in terms of resources: LUT, Slice registers, and DSP. Finally, in this article, we demonstrate the advantages of hardware implementation over software implementation in terms of power consumption, execution time and resources used, which is crucial for wireless network applications. In future research, we aim to investigate how Generative Adversarial Networks (GANs) can generate optimized path planning in wireless sensor networks (WSNs), enhancing data collection, coverage, and energy efficiency. The utilization of GANs for trajectory planning holds great potential to optimize autonomous navigation, considering factors like node density and environmental dynamics, thereby improving node location accuracy in real-world applications. CRediT authorship contribution statement Rahma Mani: Writing – original draft, Visualization, Software, Resources, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Antonio Rios-Navarro: Review & editing, Validation, Supervision. Jose Luis Sevillano Ramos: Review & editing, Validation, Supervision. Noureddine Liouane: Review & editing, Validation, Supervision. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Data availability Data will be made available on request. References [1] M. Majid, S. Habib, A.R. Javed, M. Rizwan, G. Srivastava, T.R. Gadekallu, J.C.-W. Lin, Applications of wireless sensor networks and internet of things frameworks in the industry revolution 4.0: A systematic literature review, Sensors 22 (2022) 2087. [2] S.J. Hayward, Kate Van Lopik, Christopher Hinde, et al., A survey of indoor location technologies, techniques and applications in industry, Internet Things (2022) 100608. [3] J.R. Lowell, Military applications of localization, tracking, and targeting, IEEE Wirel. Commun. 18 (2) (2011) 60–65. [4] J. Bravo-Arrabal, P. Zambrana, J.J. Fernandez-Lozano, J.A. Gomez-Ruiz, J.S. Barba, A. García-Cerezo, Realistic deployment of hybrid wireless sensor networks based on ZigBee and LoRa for search and rescue applications, IEEE Access 10 (2022) 64618–64637. [5] Vini Gupta, Swades De, An energy-efficient edge computing framework for decentralized sensing in WSN-assisted IoT, IEEE Trans. Wirel. Commun. 20 (8) (2021) 4811–4827. [6] De-gan Zhang, Chen-hao Ni, Jie Zhang, et al., A novel edge computing architecture based on adaptive stratified sampling, Comput. Commun. 183 (2022) 121–135. [7] Tian Wang, Yucheng Lu, Zhihan Cao, et al., When sensor-cloud meets mobile edge computing, Sensors 19 (23) (2019) 5324. [8] Inam Ullah, Siyu Qian, Zhixiang Deng, et al., Extended Kalman filter-based localization algorithm by edge computing in wireless sensor networks, Digit. Commun. Netw. 7 (2) (2021) 187–195. [9] Zeyu Sun, Lili Wei, Chen Xu, et al., An energy-efficient cross-layer-sensing clustering method based on intelligent fog computing in WSNs, IEEE Access 7 (2019) 144165-144177. Pervasive and Mobile Computing 103 (2024) 101961 17 R. Mani et al. [10] R. Mani, S. Messous, G. Campobello F. Battaglia, N. Liouane, Improved distance vector based kalman filter localization algorithm for wireless sensor network, in: 2023 International Conference on Control, Automation and Diagnosis, (ICCAD), Rome, Italy, 2023. [11] Elhadi Shakshukia, Abdulrahman Abou Elkhailb, Ibrahim Nemerb, Mumin Adamb, Tarek Sheltamib, Comparative study on range free localization algorithms, in: The 10th International Conference on Ambient Systems, Networks and Technologies, (ANT), April 29 - May 2, Leuven, Belgium, 2019. [12] R. Mani, J.L. Sevillano, N. Liouane, Improved least-square dv-hop algorithm for localization in large scale wireless sensor network, in: 2022 International Conference on Control, Automation and Diagnosis, (ICCAD), Lisbon, Portugal, 2022. [13] P. Huang, L. Zeng, X. Chen, K. Luo, Z. Zhou, S. Yu, Edge robotics: Edge-computing-accelerated multirobot simultaneous localization and mapping, IEEE Internet Things J. 9 (15) (2022) 14087–14102, 1 Aug.1. [14] O. Azbar, B. Ors, G.K. Kurt, Implementation of an indoor localization algorithms on an fpga, in: 2015 9th International Conference on Electrical and Electronics Engineering, (ELECO), Bursa, Turkey, 2015, pp. 949–952. [15] A. Ghorbel, N. Ben Amor, M. Jallouli, L. Amouri, A HW/SW implementation on FPGA of a robot localization algorithm, in: International Multi-Conference on Systems, Signals & Devices, Chemnitz, Germany, 2012, pp. 1–7. [16] P. Benoit, L. Dalmasso, G. Patrigeon, T. Gil, F. Bruguier, L. Torres, Edge-computing perspectives with reconfigurable hardware, in: 2019 14th International Symposium on Reconfigurable Communication-centric Systems-on-Chip (ReCoSoC), York, UK, 2019, pp. 51–58. [17] R. Ferdian, R. Aisuwarya, T. Erlina, Edge computing for internet of things based on FPGA, in: 2020 International Conference on Information Technology Systems and Innovation, (ICITSI), Bandung, Indonesia, 2020, pp. 20–23. [18] Milan Groshev, Gabriele Baldoni, Luca Cominardi, Antonio de la Oliva, Robert Gazda, Edge robotics: are we ready? an experimental evaluation of current vision and future directions, Digit. Commun. Netw. 9 (1) (2023) 166–174. [19] Ibrahim M. Zeki, Mohammed A. Hashim, Maytham M. Hammood, Rabah Nori, A review on localization algorithms of mobile robot in different environments, J. Algebr. Stat. 13 (2) (2022). [20] A. Ghorbel, N. Ben Amor, M. Jallouli, Design of a flexible reconfigurable mobile robot localization system using FPGA technology, SN Appl. Sci. 2 (2020) 1183. [21] L.L.D. Oliveira, G.H. Eisenkraemer, E.A. Carara, J.B. Martins, J. Monteiro, Mobile localization techniques for wireless sensor networks: Survey and recommendations, ACM Trans. Sensor Netw. 19 (2) (2023) 1–39. [22] Y. Emami, B. Wei, K. Li, W. Ni, E. Tovar, Joint communication scheduling and velocity control in multi-UAV-assisted sensor networks: A deep reinforcement learning approach, IEEE Trans. Veh. Technol. 70 (10) (2021) 10986–10998. [23] T. Li, C. Wang, Q. Na, Research on DV-hop improved algorithm based on dual communication radius, J. Wirel. Com Netw. 2020 (2020) 113. [24] R. Mani, A. Rios-Navarro, J.L. Sevillano-Ramos, et al., Improved 3D localization algorithm for large scale wireless sensor networks, Wirel. Netw. (2023). [25] A. Ghorbel, N. Ben Amor, M. Jallouli, Design of a flexible reconfigurable mobile robot localization system using FPGA technology, SN Appl. Sci. 2 (2020) 1183. [26] A. Irwansyah, O.W. Ibraheem, J. Hagemeyer, M. Porrmann, U. Rueckert, FPGA-based multi-robot tracking, J. Parallel Distrib. Comput. 107 (2017) 146–161. [27] K. Vennela, M.C. Chinnaaiah, S. Dubey, S. Savithri, Implementation of mobile robot navigation mechanism using fpga: an edge detection-based approach, in: S. Bhattacharyya, A. Hassanien, D. Gupta, A. Khanna, I. Pan (Eds.), International Conference on Innovative Computing and Communications, 2019. [28] Mohammed Alareqi, Rachid Elgouri, Khalid Mateur, Laamari Hlou, Real time hardware co-simulation for image processing algorithms using xilinx system generator, Int. J. Electr. Eng. Inform.. [29] R.N. Biswas, A. Saha, S.K. Mitra, et al., Design and implementation of anchor coprocessor architecture for wireless node localization applications, Peer-to-Peer Netw. Appl. (2024). [30] Y. Wu, C. Zhang, L. Tong, X. Shi, Location optimization based on improved 3D DV-HOP algorithm in wireless sensor networks, IEEE Access 11 (2023) 85525–85536. [31] M.R. Hussam, M.S. Omar, H. Noha, FPGA implementation in mobile robot applications: State of the art review, Multidiszciplináris Tudományok 13 (2) (2023) 232–249. [32] Xin Shi, Linghua Zhang, High-precision weighted bounding box localization algorithm for wireless sensor network, in: 2013 IEEE Third International Conference on Information Science and Technology, (ICIST), IEEE, 2013, pp. 1110–1113. [33] O. Liouane, S. Femmam, T. Bakir, et al., New online DV-hop algorithm via mobile anchor for wireless sensor network localization, Tsinghua Sci. Technol. 28 (5) (2023) 940–951.