Pub Date : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551109
Pengshan Chen, W. Wang, Jingjie Gao
Interference suppression using adaptive nulling algorithm is an important array signal processing technique for radar/sonar sensing. However, in long term task, most of the arrays’ parameters vary from time to time, which need known sources to re-calibrate. To be free of calibration sources, this paper presents an adaptive nulling algorithm using array observation data. We first establish the model of steering vector (SV) mismatches due to gain-phase error and sensor shifting. Then the angle-related bases of received signal subspace are estimated by applying a joint optimization method consists of Genetic algorithm (GA) and quasi-Newton method. In the end, the array weighting vector can be calculated, and the results of several numerical simulations are demonstrated, which shows that the proposed algorithm can significantly improve the interference suppression performance of sensor array.
{"title":"Interference Suppression Using Adaptive Nulling Algorithm Without Calibration Sources","authors":"Pengshan Chen, W. Wang, Jingjie Gao","doi":"10.1109/ICAS49788.2021.9551109","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551109","url":null,"abstract":"Interference suppression using adaptive nulling algorithm is an important array signal processing technique for radar/sonar sensing. However, in long term task, most of the arrays’ parameters vary from time to time, which need known sources to re-calibrate. To be free of calibration sources, this paper presents an adaptive nulling algorithm using array observation data. We first establish the model of steering vector (SV) mismatches due to gain-phase error and sensor shifting. Then the angle-related bases of received signal subspace are estimated by applying a joint optimization method consists of Genetic algorithm (GA) and quasi-Newton method. In the end, the array weighting vector can be calculated, and the results of several numerical simulations are demonstrated, which shows that the proposed algorithm can significantly improve the interference suppression performance of sensor array.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130588829","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 : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551147
Gabriela Gresenz, Jules White, D. Schmidt
This paper describes the structure and functionality of a dataset designed to enable autonomous vehicles to learn about off-road terrain using a single monocular image. This dataset includes over 12,000 images of off-road terrain and the corresponding sensor data from a global positioning system (GPS), inertial measurement units (IMUs), and a wheel rotation speed sensor. The paper also describes and empirically evaluates eight roughness labeling schemas derived from IMU z-axis acceleration for labeling the images in our dataset. These roughness labels can be used for training deep learning models to detect terrain roughness.
{"title":"An Off-Road Terrain Dataset Including Images Labeled With Measures Of Terrain Roughness","authors":"Gabriela Gresenz, Jules White, D. Schmidt","doi":"10.1109/ICAS49788.2021.9551147","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551147","url":null,"abstract":"This paper describes the structure and functionality of a dataset designed to enable autonomous vehicles to learn about off-road terrain using a single monocular image. This dataset includes over 12,000 images of off-road terrain and the corresponding sensor data from a global positioning system (GPS), inertial measurement units (IMUs), and a wheel rotation speed sensor. The paper also describes and empirically evaluates eight roughness labeling schemas derived from IMU z-axis acceleration for labeling the images in our dataset. These roughness labels can be used for training deep learning models to detect terrain roughness.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"33 6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124410079","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 : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551111
Giulia Slavic, P. Marín, David Martín, L. Marcenaro, C. Regazzoni
When performing anomaly detection on an autonomous vehicle’s sensory data, it is fundamental to infer the cause of the found anomalies. This paper proposes a method for learning prediction models and detecting anomalies by decomposing the evolution of an agent’s state into its different motion-related parameters. A filter is introduced based on Generalized Filtering to increase the interpretability of the results with respect to previous methods. The proposed anomaly detection method is tested on data from a real vehicle. We also consider the case in which multiple models are learned, how to extract the salient discriminatory features of each, and use the proposed anomaly detection method to perform behavior classification.
{"title":"Interpretable Anomaly Detection Using A Generalized Markov Jump Particle Filter","authors":"Giulia Slavic, P. Marín, David Martín, L. Marcenaro, C. Regazzoni","doi":"10.1109/ICAS49788.2021.9551111","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551111","url":null,"abstract":"When performing anomaly detection on an autonomous vehicle’s sensory data, it is fundamental to infer the cause of the found anomalies. This paper proposes a method for learning prediction models and detecting anomalies by decomposing the evolution of an agent’s state into its different motion-related parameters. A filter is introduced based on Generalized Filtering to increase the interpretability of the results with respect to previous methods. The proposed anomaly detection method is tested on data from a real vehicle. We also consider the case in which multiple models are learned, how to extract the salient discriminatory features of each, and use the proposed anomaly detection method to perform behavior classification.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"78 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127594964","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 : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551193
Alexander Schmidt, Walter Kellermann
The suppression of ego-noise is often addressed using dictionary-based methods where the characteristic spectral structure of ego-noise is approximated by a linear combination of dictionary entries. A blind, entirely audio data-based selection of the dictionary entries is, however, challenging and reacts sensitive against other signals besides ego-noise in a mixture. For a more robust behavior, we propose a motor data-dependent regularization term which promotes similar activations for similar physical states of the robot. The proposed regularization term is added to a multichannel nonnegative matrix factorization (MNMF)-based signal model and according update rules are derived. We analyze the proposed method for a challenging ego-noise scenario and demonstrate the efficacy of the method compared to an approach for which no motor data is used.
{"title":"Multichannel Nonnegative Matrix Factorization With Motor Data-Regularized Activations For Robust Ego-Noise Suppression","authors":"Alexander Schmidt, Walter Kellermann","doi":"10.1109/ICAS49788.2021.9551193","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551193","url":null,"abstract":"The suppression of ego-noise is often addressed using dictionary-based methods where the characteristic spectral structure of ego-noise is approximated by a linear combination of dictionary entries. A blind, entirely audio data-based selection of the dictionary entries is, however, challenging and reacts sensitive against other signals besides ego-noise in a mixture. For a more robust behavior, we propose a motor data-dependent regularization term which promotes similar activations for similar physical states of the robot. The proposed regularization term is added to a multichannel nonnegative matrix factorization (MNMF)-based signal model and according update rules are derived. We analyze the proposed method for a challenging ego-noise scenario and demonstrate the efficacy of the method compared to an approach for which no motor data is used.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124098488","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 : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551201
Siwei Zhang, E. Staudinger, R. Pöhlmann, A. Dammann
Networks composed of a myriad of autonomous robots have attracted increasing attention in recent years due to their enormous capability expansion from single robot systems. In these networks, robots benefit from the collaboration with each other to enhance their situation awareness for autonomous operation. For example, in an extraterrestrial exploration mission, a robotic swarm can collaboratively utilize the inter-robot communication system to propagate information, synchronize itself, and navigate to achieve mission objectives like joint environmental sensing. In addition, each robot can decide and control its own trajectory, so that the aforementioned tasks are accomplished in a globally efficient manner. In this paper, we propose multi-agent control strategies for autonomous robotic networks, which adapt the mission demands on cooperative communication, localization and sensing. We also discuss three space exploration examples with different mission demands, which lead to distinct network formations. These three missions will be conceptually demonstrated in a space analog mission on the volcano Mount Etna in June 2022.
{"title":"Cooperative Communication, Localization, Sensing and Control for Autonomous Robotic Networks","authors":"Siwei Zhang, E. Staudinger, R. Pöhlmann, A. Dammann","doi":"10.1109/ICAS49788.2021.9551201","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551201","url":null,"abstract":"Networks composed of a myriad of autonomous robots have attracted increasing attention in recent years due to their enormous capability expansion from single robot systems. In these networks, robots benefit from the collaboration with each other to enhance their situation awareness for autonomous operation. For example, in an extraterrestrial exploration mission, a robotic swarm can collaboratively utilize the inter-robot communication system to propagate information, synchronize itself, and navigate to achieve mission objectives like joint environmental sensing. In addition, each robot can decide and control its own trajectory, so that the aforementioned tasks are accomplished in a globally efficient manner. In this paper, we propose multi-agent control strategies for autonomous robotic networks, which adapt the mission demands on cooperative communication, localization and sensing. We also discuss three space exploration examples with different mission demands, which lead to distinct network formations. These three missions will be conceptually demonstrated in a space analog mission on the volcano Mount Etna in June 2022.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125316590","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}
Formation and collision avoidance abilities are essential for multi-agent systems. Conventional methods usually require a central controller and global information to achieve collaboration, which is impractical in an unknown environment. In this paper, we propose a deep reinforcement learning (DRL) based distributed formation control scheme for autonomous vehicles. A modified stream-based obstacle avoidance method is applied to smoothen the optimal trajectory, and onboard sensors such as Lidar and antenna arrays are used to obtain local relative distance and angle information. The proposed scheme obtains a scalable distributed control policy which jointly optimizes formation tracking error and average collision rate with local observations. Simulation results demonstrate that our method outperforms two other state-of-the-art algorithms on maintaining formation and collision avoidance.
{"title":"A DRL Based Distributed Formation Control Scheme with Stream-Based Collision Avoidance","authors":"Xinyou Qiu, Xiaoxiang Li, Jian Wang, Yu Wang, Yuan Shen","doi":"10.1109/ICAS49788.2021.9551123","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551123","url":null,"abstract":"Formation and collision avoidance abilities are essential for multi-agent systems. Conventional methods usually require a central controller and global information to achieve collaboration, which is impractical in an unknown environment. In this paper, we propose a deep reinforcement learning (DRL) based distributed formation control scheme for autonomous vehicles. A modified stream-based obstacle avoidance method is applied to smoothen the optimal trajectory, and onboard sensors such as Lidar and antenna arrays are used to obtain local relative distance and angle information. The proposed scheme obtains a scalable distributed control policy which jointly optimizes formation tracking error and average collision rate with local observations. Simulation results demonstrate that our method outperforms two other state-of-the-art algorithms on maintaining formation and collision avoidance.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"18 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126855421","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 : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551118
Aarti Singh, Neal Patwari, McKelvey
Real-time proximity and collision detection via radio frequency (RF) distance measurements has application in smart helmets, drones, autonomous vehicles, and social distancing. In this paper we evaluate ACED, a range-based, infrastructure-free, distributed algorithm that utilizes inter-node range data and intra-node acceleration data to estimate the recent relative positions of each node and to predict impending collisions between any pair of nodes. The framework is tested and validated using experimental data from a testbed of mobile nodes which use ultra-wideband (UWB) ranging and inertial sensing. ACED is shown to outperform two state-of-the-art methods.
{"title":"Collision Prediction using UWB and Inertial Sensing: Experimental Evaluation","authors":"Aarti Singh, Neal Patwari, McKelvey","doi":"10.1109/ICAS49788.2021.9551118","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551118","url":null,"abstract":"Real-time proximity and collision detection via radio frequency (RF) distance measurements has application in smart helmets, drones, autonomous vehicles, and social distancing. In this paper we evaluate ACED, a range-based, infrastructure-free, distributed algorithm that utilizes inter-node range data and intra-node acceleration data to estimate the recent relative positions of each node and to predict impending collisions between any pair of nodes. The framework is tested and validated using experimental data from a testbed of mobile nodes which use ultra-wideband (UWB) ranging and inertial sensing. ACED is shown to outperform two state-of-the-art methods.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"26 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127583464","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 : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551198
I. Pitas
1D, 2D and multidimensional convolutions are basic tools in deep learning, notably in convolutional neural networks (CNNs) and in computer vision (template matching, correlation trackers). Therefore, fast 1D/2D/3D convolution algorithms are essential for advanced machine learning and computer vision. This paper presents: 1) novel optimal n-D cyclic convolution algorithms having minimal multiplicative complexity that are much faster than any competing convolution algorithm internationally and 2) methods for speeding up such optimal convolution algorithms on GPUs and multicore CPUs. Such a speedup is very important both for CNN training and CNN testing, particularly in embedded environments (e.g., on drones) and real-time applications (e.g., fast CNN inference for object detection and correlation trackers for embedded real-time object tracking).
{"title":"Optimal Multidimensional Cyclic Convolution Algorithms For Deep Learning And Computer Vision Applications","authors":"I. Pitas","doi":"10.1109/ICAS49788.2021.9551198","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551198","url":null,"abstract":"1D, 2D and multidimensional convolutions are basic tools in deep learning, notably in convolutional neural networks (CNNs) and in computer vision (template matching, correlation trackers). Therefore, fast 1D/2D/3D convolution algorithms are essential for advanced machine learning and computer vision. This paper presents: 1) novel optimal n-D cyclic convolution algorithms having minimal multiplicative complexity that are much faster than any competing convolution algorithm internationally and 2) methods for speeding up such optimal convolution algorithms on GPUs and multicore CPUs. Such a speedup is very important both for CNN training and CNN testing, particularly in embedded environments (e.g., on drones) and real-time applications (e.g., fast CNN inference for object detection and correlation trackers for embedded real-time object tracking).","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121568921","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 : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551179
Yixiao Zhang, Gang Chen, Tingting Zhang
Since multiple roads merge at intersections, proper coordination for vehicles is of great importance for modern intelligent transportation systems (ITS). In this paper, we try to smartly integrate the infrastructure and vehicle-based planners, to achieve feasible and efficient solutions. In detail, the vehicle reference trajectories can be firstly achieved by the high-level infrastructure-based coordination, which can be formulated as standard quadratic programming (QP) and mixed integer programming (MIP) problems. Due to the possible occurrence of obstacles such as pedestrians, the vehicles are also required to perform low-level ego trajectory optimization based on local observations, which are essentially dynamic programming (DP) and QP problems. Numerical results show that the proposed framework can effectively solve many opening problems in vehicle coordination, such as obstacle avoidance and deadlocks among vehicles.
{"title":"Intelligent Intersection Coordination and Trajectory Optimization for Autonomous Vehicles","authors":"Yixiao Zhang, Gang Chen, Tingting Zhang","doi":"10.1109/ICAS49788.2021.9551179","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551179","url":null,"abstract":"Since multiple roads merge at intersections, proper coordination for vehicles is of great importance for modern intelligent transportation systems (ITS). In this paper, we try to smartly integrate the infrastructure and vehicle-based planners, to achieve feasible and efficient solutions. In detail, the vehicle reference trajectories can be firstly achieved by the high-level infrastructure-based coordination, which can be formulated as standard quadratic programming (QP) and mixed integer programming (MIP) problems. Due to the possible occurrence of obstacles such as pedestrians, the vehicles are also required to perform low-level ego trajectory optimization based on local observations, which are essentially dynamic programming (DP) and QP problems. Numerical results show that the proposed framework can effectively solve many opening problems in vehicle coordination, such as obstacle avoidance and deadlocks among vehicles.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"211 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123390092","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 : 2021-08-11DOI: 10.1109/ICAS49788.2021.9551189
Neeraj Pandey, S. S. Ram
Automated driving tests using cameras have been researched for expediting the training and testing of car drivers. We propose an automated parking test using millimeter-wave automotive radars. The advantage is that these radars can be operated even in low visibility conditions. We propose generating high-resolution inverse synthetic aperture radar (ISAR) images of a vehicle under test (VUT) parking into a designated parking slot from an externally mounted radar. The trajectory of the motion is estimated from the ISAR data using polynomial curve fitting from which the VUT is deemed to have either correctly or incorrectly parked. We experimentally validate the proposed method with millimeter-wave radar data gathered for cars performing perpendicular and 45° angle parking.
{"title":"Automated Parking Test Using ISAR Images from Automotive Radar","authors":"Neeraj Pandey, S. S. Ram","doi":"10.1109/ICAS49788.2021.9551189","DOIUrl":"https://doi.org/10.1109/ICAS49788.2021.9551189","url":null,"abstract":"Automated driving tests using cameras have been researched for expediting the training and testing of car drivers. We propose an automated parking test using millimeter-wave automotive radars. The advantage is that these radars can be operated even in low visibility conditions. We propose generating high-resolution inverse synthetic aperture radar (ISAR) images of a vehicle under test (VUT) parking into a designated parking slot from an externally mounted radar. The trajectory of the motion is estimated from the ISAR data using polynomial curve fitting from which the VUT is deemed to have either correctly or incorrectly parked. We experimentally validate the proposed method with millimeter-wave radar data gathered for cars performing perpendicular and 45° angle parking.","PeriodicalId":287105,"journal":{"name":"2021 IEEE International Conference on Autonomous Systems (ICAS)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2021-08-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114954307","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：