Pub Date : 2023-02-10DOI: 10.1109/ICARA56516.2023.10125619
Danial Pour Arab, Matthias Spisser, C. Essert
Over the past two decades, an evolutionary effort has been established in the agricultural sector to develop efficient autonomous systems that can carry out common in-field operations including harvesting, mowing, and spraying. Increasing production while decreasing costs and environmental damages is one of the main objectives for these autonomous systems. Due to the nature of these tasks, complete coverage path planning techniques are crucial to determining the best path that covers the entire field while accounting for terrain characteristics, operational needs, and robot properties. In this study, we propose a novel complete coverage path planning approach to define the ideal path for a wheeled robot across an agricultural field. To identify all feasible solutions satisfying a set of predefined constraints, a method based on tree exploration is first proposed that examines row-skip patterns. Second, the most optimal solution is selected by a selection method. Maximizing the covered area while minimizing overlaps, non-working path length, number of turns containing reverse moves, and overall travel time are the objectives of the selection method. We showed on 6 real-world fields geometries that the row skip approach offered benefits in terms of reduction of the required headland size, and often helped decreasing the number of necessary reverse moves and the overlaps, while increasing the covered area.
{"title":"Introduction of A Row-Skip Pattern in Complete Coverage Path Planning for Agricultural Fields","authors":"Danial Pour Arab, Matthias Spisser, C. Essert","doi":"10.1109/ICARA56516.2023.10125619","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125619","url":null,"abstract":"Over the past two decades, an evolutionary effort has been established in the agricultural sector to develop efficient autonomous systems that can carry out common in-field operations including harvesting, mowing, and spraying. Increasing production while decreasing costs and environmental damages is one of the main objectives for these autonomous systems. Due to the nature of these tasks, complete coverage path planning techniques are crucial to determining the best path that covers the entire field while accounting for terrain characteristics, operational needs, and robot properties. In this study, we propose a novel complete coverage path planning approach to define the ideal path for a wheeled robot across an agricultural field. To identify all feasible solutions satisfying a set of predefined constraints, a method based on tree exploration is first proposed that examines row-skip patterns. Second, the most optimal solution is selected by a selection method. Maximizing the covered area while minimizing overlaps, non-working path length, number of turns containing reverse moves, and overall travel time are the objectives of the selection method. We showed on 6 real-world fields geometries that the row skip approach offered benefits in terms of reduction of the required headland size, and often helped decreasing the number of necessary reverse moves and the overlaps, while increasing the covered area.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"443 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133213217","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-10DOI: 10.1109/ICARA56516.2023.10125740
Alexander Poeppel, Christian Eymüller, W. Reif
Human-robot-collaboration (HRC) requires fast and reliable sensor data to ensure the safety of humans in the workspace. Current solutions for processing multi-modal sensor data in HRC are either highly performant in specific scenarios or offer more flexibility at the cost of decreased performance. Our GPU accelerated SensorClouds framework, however, combines both high flexibility and real-time performance. The architecture aids developers in quickly implementing complex HRC applications with multiple sensors by encapsulating all functionality into reusable modules. The resulting pipeline is optimized by the framework and executed in real-time.
{"title":"SensorClouds: A Framework for Real-Time Processing of Multi-modal Sensor Data for Human-Robot-Collaboration","authors":"Alexander Poeppel, Christian Eymüller, W. Reif","doi":"10.1109/ICARA56516.2023.10125740","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125740","url":null,"abstract":"Human-robot-collaboration (HRC) requires fast and reliable sensor data to ensure the safety of humans in the workspace. Current solutions for processing multi-modal sensor data in HRC are either highly performant in specific scenarios or offer more flexibility at the cost of decreased performance. Our GPU accelerated SensorClouds framework, however, combines both high flexibility and real-time performance. The architecture aids developers in quickly implementing complex HRC applications with multiple sensors by encapsulating all functionality into reusable modules. The resulting pipeline is optimized by the framework and executed in real-time.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114753301","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-10DOI: 10.1109/ICARA56516.2023.10125613
Oguz Kedilioglu, M. Nikol, J. Walter, J. Franke
The motion planning accuracy of articulated robots can be increased by augmenting them with an objective function. Parameters that influence the absolute accuracy of 6-axis articulated robots can be integrated into such objective functions. However, it is still unclear how to prioritize the various influencing factors so that they can be utilized together instead of separately. Here, it is shown how to rate these factors by analyzing and comparing their impact on the accuracy of the robot. The factors manipulability, smoothness, gear backlash caused by direction change of axis rotation, joint space distance and robot layout are studied in this regard. The evaluation with a 6DoF laser tracker reveals that manipulability and smoothness have the biggest impact. With these results it is possible to combine multiple factors into a comprehensive function and achieve a higher accuracy without the necessity of applying any hardware upgrades.
{"title":"Correlation Analysis of Factors Influencing the Motion Planning Accuracy of Articulated Robots","authors":"Oguz Kedilioglu, M. Nikol, J. Walter, J. Franke","doi":"10.1109/ICARA56516.2023.10125613","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125613","url":null,"abstract":"The motion planning accuracy of articulated robots can be increased by augmenting them with an objective function. Parameters that influence the absolute accuracy of 6-axis articulated robots can be integrated into such objective functions. However, it is still unclear how to prioritize the various influencing factors so that they can be utilized together instead of separately. Here, it is shown how to rate these factors by analyzing and comparing their impact on the accuracy of the robot. The factors manipulability, smoothness, gear backlash caused by direction change of axis rotation, joint space distance and robot layout are studied in this regard. The evaluation with a 6DoF laser tracker reveals that manipulability and smoothness have the biggest impact. With these results it is possible to combine multiple factors into a comprehensive function and achieve a higher accuracy without the necessity of applying any hardware upgrades.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"72 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115348897","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-10DOI: 10.1109/ICARA56516.2023.10125681
Jianyu Tang, Tao Zhou, E. Zakeri, Tingting Shu, W. Xie
This paper proposes a novel accurate dynamic path tracking (DPT) method for industrial robots based on photogrammetry sensors and an adaptive neuro-PID (ANPID) control method. First, the pose of the robot's end-effector is detected by the photogrammetry sensor (C-Track stereo camera). It passes through a robust Kalman filter to reduce the noise in the signals. Then, the filtered signals are fed to the ANPID, whose gains are tuned online using an adaptive multi-layer perceptron neural network (AMLPNN). The steepest descent optimization method is adopted online. The cost function is the least mean square of the system states errors. Experimental results on FANUC M-20iA robot show the tracking accuracy reaches ±0.08mm and ±0.04deg, which exhibits the superiority of the proposed method over the conventional methods such as PID (tracking error±0.2mm and ±0.1deg) [4].
{"title":"Photogrammetry-based Dynamic Path Tracking of Industrial Robots Using Adaptive Neuro-PID Control Method and Robust Kalman Filter","authors":"Jianyu Tang, Tao Zhou, E. Zakeri, Tingting Shu, W. Xie","doi":"10.1109/ICARA56516.2023.10125681","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125681","url":null,"abstract":"This paper proposes a novel accurate dynamic path tracking (DPT) method for industrial robots based on photogrammetry sensors and an adaptive neuro-PID (ANPID) control method. First, the pose of the robot's end-effector is detected by the photogrammetry sensor (C-Track stereo camera). It passes through a robust Kalman filter to reduce the noise in the signals. Then, the filtered signals are fed to the ANPID, whose gains are tuned online using an adaptive multi-layer perceptron neural network (AMLPNN). The steepest descent optimization method is adopted online. The cost function is the least mean square of the system states errors. Experimental results on FANUC M-20iA robot show the tracking accuracy reaches ±0.08mm and ±0.04deg, which exhibits the superiority of the proposed method over the conventional methods such as PID (tracking error±0.2mm and ±0.1deg) [4].","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"265 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122907757","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-10DOI: 10.1109/ICARA56516.2023.10125724
Mia Krantz, Niklas Widulle, O. Niggemann
The high complexity of Cyber-Physical Systems (CPS) necessitates novel approaches for system analysis, planning, anomaly detection and testing. Machine Learning (ML) methods are promising because of their ability to find underlying relations even in large, complex and conflicting data. While existing CPS produce large data sets, these might not cover the appropriate time frame, or the desired configuration. Therefore, the use of ML methods requires the use of simulation tools to generate the necessary data. There are numerous approaches to simulate CPS. However, they often have significant shortcomings regarding their expressiveness in regards to physical properties of system components, their scalability in the face of the ever-increasing complexity of CPS, their usability for simultaneous simulation of different aspects of CPS and interoperability between different simulation environments. Game and media creation tools have seen an impressive development in recent years with regards to their realistic representation of physical systems and simulation capabilities. These are already employed in some engineering challenges like training of algorithms for self-driving cars. They have huge potential for the application in simulation and analysis of CPS. In this work we provide an analysis of the shortcomings of currently used environments for modeling and simulation of CPS with regards to creating data for ML. We then analyze how currently existing limitations can be overcome by employing tools from game and media design, discussing possible use cases and applications of these tools. With this, we present a possible new direction of research which has the potential to improve modeling of CPS, especially with regards to their application for ML.
{"title":"Game Design Tools for ML Data Generation in CPS","authors":"Mia Krantz, Niklas Widulle, O. Niggemann","doi":"10.1109/ICARA56516.2023.10125724","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125724","url":null,"abstract":"The high complexity of Cyber-Physical Systems (CPS) necessitates novel approaches for system analysis, planning, anomaly detection and testing. Machine Learning (ML) methods are promising because of their ability to find underlying relations even in large, complex and conflicting data. While existing CPS produce large data sets, these might not cover the appropriate time frame, or the desired configuration. Therefore, the use of ML methods requires the use of simulation tools to generate the necessary data. There are numerous approaches to simulate CPS. However, they often have significant shortcomings regarding their expressiveness in regards to physical properties of system components, their scalability in the face of the ever-increasing complexity of CPS, their usability for simultaneous simulation of different aspects of CPS and interoperability between different simulation environments. Game and media creation tools have seen an impressive development in recent years with regards to their realistic representation of physical systems and simulation capabilities. These are already employed in some engineering challenges like training of algorithms for self-driving cars. They have huge potential for the application in simulation and analysis of CPS. In this work we provide an analysis of the shortcomings of currently used environments for modeling and simulation of CPS with regards to creating data for ML. We then analyze how currently existing limitations can be overcome by employing tools from game and media design, discussing possible use cases and applications of these tools. With this, we present a possible new direction of research which has the potential to improve modeling of CPS, especially with regards to their application for ML.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129952348","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-10DOI: 10.1109/ICARA56516.2023.10125728
Samia Choueiri, H. Diab, M. Owayjan, Roger Achkar
Humanoid robots are capable of imitating humans, adapting to changes in an environment, making decisions, and performing tasks. But they are also at a high risk of being hit by an external impact force or disturbance. However, unlike humans, robots do not have sensory neurons to be able to sense the location at which they were hit as well as the intensity of the impact force. In this paper, humanoids are tested for the ability to perceive their surrounding better by recognizing the impact force location and intensity after monitoring the response of the encoders at the different joints due to an external disturbance where additional sensors to the robot are not needed. Using machine learning, several models are trained and then tested to ameliorate and increase the robustness of the stability control algorithm as it can help the robot regain stability in a more educated system. Giving the humanoid the ability to perceive its surroundings better also gives it the ability to react and control its movement and posture with a better and faster response despite the increasing number of degrees of freedom which makes controlling the robot more difficult.
{"title":"Impact Force Location and Intensity Identification Using Joint-Position Sensors in Humanoids","authors":"Samia Choueiri, H. Diab, M. Owayjan, Roger Achkar","doi":"10.1109/ICARA56516.2023.10125728","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125728","url":null,"abstract":"Humanoid robots are capable of imitating humans, adapting to changes in an environment, making decisions, and performing tasks. But they are also at a high risk of being hit by an external impact force or disturbance. However, unlike humans, robots do not have sensory neurons to be able to sense the location at which they were hit as well as the intensity of the impact force. In this paper, humanoids are tested for the ability to perceive their surrounding better by recognizing the impact force location and intensity after monitoring the response of the encoders at the different joints due to an external disturbance where additional sensors to the robot are not needed. Using machine learning, several models are trained and then tested to ameliorate and increase the robustness of the stability control algorithm as it can help the robot regain stability in a more educated system. Giving the humanoid the ability to perceive its surroundings better also gives it the ability to react and control its movement and posture with a better and faster response despite the increasing number of degrees of freedom which makes controlling the robot more difficult.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"84 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126081730","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-02-10DOI: 10.1109/ICARA56516.2023.10125773
Ammar K. Al Mhdawi, N. Wright, S. Benson, M. Haroutunian
In nuclear inspection environments, a tether cable is used to transfer power and data between the underwater robotic system and the surface control unit. During underwater nuclear inspection, the tether cable can become entangled and loop with the environment such as nuclear waste boxes and objects. The risk of colliding with underwater objects is increased by the presence of more inspection robots underwater, especially if they are equipped with manipulator arms. As a result of the loops and knots around the cable, the inspection process may be affected and the ROV may not be able to perform its job. The present article is an extended development of the previous Collision Avoidance Robotic Tether (CART-I) model [1]. The CART-I system consists of micro thrusters that are attached to the base unit by a tether cable. The micro thrust unit is capable of generating a small amount of thrust that can move the tether away from obstacles in the water, particularly in restricted spaces. The use of light detection technologies such as IR or LiDAR for obstacle detection is not effective underwater due to the complex motion dynamics of the tether underwater and the size of obstacles, which makes it impossible to provide definite identification of the objects within a given time period. In order to provide the surroundings of the micro thrust units with obstacle detection capability, we have developed an autonomous force soft sensor. Additionally, the soft moulded sealed encase was developed for effective force detection underwater, and was experimentally tested in a water tank to validate our proposed design. Simulation and experimental results of the sensor is provided. The overall goal of the CART-II is to provide a smart autonomous vision of obstacle avoidance using soft force sensing capabilities. This paper presents the full kinematic model and the simulation with finite element analysis of the CART-II system with the hardware and physical implementation of the soft sensor in order to enhance the performance of traditional tether systems.
{"title":"CART-II: Development of Collision Avoidance Robotic Tether with Soft Sensing Capabilities for Underwater Nuclear Inspection Vehicles","authors":"Ammar K. Al Mhdawi, N. Wright, S. Benson, M. Haroutunian","doi":"10.1109/ICARA56516.2023.10125773","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125773","url":null,"abstract":"In nuclear inspection environments, a tether cable is used to transfer power and data between the underwater robotic system and the surface control unit. During underwater nuclear inspection, the tether cable can become entangled and loop with the environment such as nuclear waste boxes and objects. The risk of colliding with underwater objects is increased by the presence of more inspection robots underwater, especially if they are equipped with manipulator arms. As a result of the loops and knots around the cable, the inspection process may be affected and the ROV may not be able to perform its job. The present article is an extended development of the previous Collision Avoidance Robotic Tether (CART-I) model [1]. The CART-I system consists of micro thrusters that are attached to the base unit by a tether cable. The micro thrust unit is capable of generating a small amount of thrust that can move the tether away from obstacles in the water, particularly in restricted spaces. The use of light detection technologies such as IR or LiDAR for obstacle detection is not effective underwater due to the complex motion dynamics of the tether underwater and the size of obstacles, which makes it impossible to provide definite identification of the objects within a given time period. In order to provide the surroundings of the micro thrust units with obstacle detection capability, we have developed an autonomous force soft sensor. Additionally, the soft moulded sealed encase was developed for effective force detection underwater, and was experimentally tested in a water tank to validate our proposed design. Simulation and experimental results of the sensor is provided. The overall goal of the CART-II is to provide a smart autonomous vision of obstacle avoidance using soft force sensing capabilities. This paper presents the full kinematic model and the simulation with finite element analysis of the CART-II system with the hardware and physical implementation of the soft sensor in order to enhance the performance of traditional tether systems.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-02-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121894704","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-01-06DOI: 10.1109/ICARA56516.2023.10125597
Ryan Donald, Peter Gavriel, Adam Norton, S. Ahmadzadeh
In this paper we focus on the evaluation of contextual autonomy for robots. More specifically, we propose a fuzzy framework for calculating the autonomy score for a small Unmanned Aerial Systems (sUAS) for performing a task while considering task complexity and environmental factors. Our framework is a cascaded Fuzzy Inference System (cFIS) composed of combination of three FIS which represent different contextual autonomy capabilities. We performed several experiments to test our framework in various contexts, such as endurance time, navigation, take off/land, and room clearing, with seven different sUAS. We introduce a predictive measure which improves upon previous predictive measures, allowing for previous real-world task performance to be used in predicting future mission performance.
{"title":"Contextual Autonomy Evaluation of Unmanned Aerial Vehicles in Subterranean Environments","authors":"Ryan Donald, Peter Gavriel, Adam Norton, S. Ahmadzadeh","doi":"10.1109/ICARA56516.2023.10125597","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125597","url":null,"abstract":"In this paper we focus on the evaluation of contextual autonomy for robots. More specifically, we propose a fuzzy framework for calculating the autonomy score for a small Unmanned Aerial Systems (sUAS) for performing a task while considering task complexity and environmental factors. Our framework is a cascaded Fuzzy Inference System (cFIS) composed of combination of three FIS which represent different contextual autonomy capabilities. We performed several experiments to test our framework in various contexts, such as endurance time, navigation, take off/land, and room clearing, with seven different sUAS. We introduce a predictive measure which improves upon previous predictive measures, allowing for previous real-world task performance to be used in predicting future mission performance.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-01-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126673880","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-24DOI: 10.1109/ICARA56516.2023.10125781
Rachmad Vidya Wicaksana Putra, Muhammad Shafique
Autonomous mobile agents such as unmanned aerial vehicles (UAVs) and mobile robots have shown huge potential for improving human productivity. These mobile agents require low power/energy consumption to have a long lifespan since they are usually powered by batteries. These agents also need to adapt to changing/dynamic environments, especially when deployed in far or dangerous locations, thus requiring efficient online learning capabilities. These requirements can be fulfilled by employing Spiking Neural Networks (SNNs) since SNNs offer low power/energy consumption due to sparse computations and efficient online learning due to bio-inspired learning mechanisms. However, a methodology is still required to employ appropriate SNN models on autonomous mobile agents. Towards this, we propose a Mantis methodology to systematically employ SNNs on autonomous mobile agents to enable energy-efficient processing and adaptive capabilities in dynamic environments. The key ideas of our Mantis include the optimization of SNN operations, the employment of a bio-plausible online learning mechanism, and the SNN model selection. The experimental results demonstrate that our methodology maintains high accuracy with a significantly smaller memory footprint and energy consumption (i.e., 3.32x memory reduction and 2.9x energy saving for an SNN model with 8-bit weights) compared to the baseline network with 32-bit weights. In this manner, our Mantis enables the employment of SNNs for resource- and energy-constrained mobile agents.
{"title":"Mantis: Enabling Energy-Efficient Autonomous Mobile Agents with Spiking Neural Networks","authors":"Rachmad Vidya Wicaksana Putra, Muhammad Shafique","doi":"10.1109/ICARA56516.2023.10125781","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125781","url":null,"abstract":"Autonomous mobile agents such as unmanned aerial vehicles (UAVs) and mobile robots have shown huge potential for improving human productivity. These mobile agents require low power/energy consumption to have a long lifespan since they are usually powered by batteries. These agents also need to adapt to changing/dynamic environments, especially when deployed in far or dangerous locations, thus requiring efficient online learning capabilities. These requirements can be fulfilled by employing Spiking Neural Networks (SNNs) since SNNs offer low power/energy consumption due to sparse computations and efficient online learning due to bio-inspired learning mechanisms. However, a methodology is still required to employ appropriate SNN models on autonomous mobile agents. Towards this, we propose a Mantis methodology to systematically employ SNNs on autonomous mobile agents to enable energy-efficient processing and adaptive capabilities in dynamic environments. The key ideas of our Mantis include the optimization of SNN operations, the employment of a bio-plausible online learning mechanism, and the SNN model selection. The experimental results demonstrate that our methodology maintains high accuracy with a significantly smaller memory footprint and energy consumption (i.e., 3.32x memory reduction and 2.9x energy saving for an SNN model with 8-bit weights) compared to the baseline network with 32-bit weights. In this manner, our Mantis enables the employment of SNNs for resource- and energy-constrained mobile agents.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133679898","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-11-24DOI: 10.1109/ICARA56516.2023.10125851
Wayne Paul Martis, Sachit Rao
In this paper, the collision avoidance problem for non-holonomic robots moving at constant linear speeds in the 2-D plane is considered. The maneuvers to avoid collisions are designed using dynamic vortex potential fields (PFs) that are functions of relative velocities in polar coordinates. Introduction of vorticity in the calculation of the gradients leads to a cooperative collision avoidance behaviour between the robots and also ensures the absence of local minima. Such a repulsive field is activated by a robot only when it is on a collision path with other mobile robots or stationary obstacles. By analysing the kinematics-based engagement dynamics in polar coordinates, the PF parameters are identified that ensure collision avoidance with stationary and moving robots, as well as those actively seeking to collide with it. Experimental results acquired using a mobile robot platform that support the theoretical contributions are presented.
{"title":"Cooperative Collision Avoidance in Mobile Robots using Dynamic Vortex Potential Fields","authors":"Wayne Paul Martis, Sachit Rao","doi":"10.1109/ICARA56516.2023.10125851","DOIUrl":"https://doi.org/10.1109/ICARA56516.2023.10125851","url":null,"abstract":"In this paper, the collision avoidance problem for non-holonomic robots moving at constant linear speeds in the 2-D plane is considered. The maneuvers to avoid collisions are designed using dynamic vortex potential fields (PFs) that are functions of relative velocities in polar coordinates. Introduction of vorticity in the calculation of the gradients leads to a cooperative collision avoidance behaviour between the robots and also ensures the absence of local minima. Such a repulsive field is activated by a robot only when it is on a collision path with other mobile robots or stationary obstacles. By analysing the kinematics-based engagement dynamics in polar coordinates, the PF parameters are identified that ensure collision avoidance with stationary and moving robots, as well as those actively seeking to collide with it. Experimental results acquired using a mobile robot platform that support the theoretical contributions are presented.","PeriodicalId":443572,"journal":{"name":"2023 9th International Conference on Automation, Robotics and Applications (ICARA)","volume":"44 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-11-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115737669","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：