Recent advancements in autonomous vehicle technologies are transforming intelligent transportation systems.Artificial intelligence enables real-time sensing,decision-making,and control on embedded platforms with impro...Recent advancements in autonomous vehicle technologies are transforming intelligent transportation systems.Artificial intelligence enables real-time sensing,decision-making,and control on embedded platforms with improved efficiency.This study presents the design and implementation of an autonomous radio-controlled(RC)vehicle prototype capable of lane line detection,obstacle avoidance,and navigation through dynamic path planning.The system integrates image processing and ultrasonic sensing,utilizing Raspberry Pi for vision-based tasks and ArduinoNano for real-time control.Lane line detection is achieved through conventional image processing techniques,providing the basis for local path generation,while traffic sign classification employs a You Only Look Once(YOLO)model optimized with TensorFlow Lite to support navigation decisions.Images captured by the onboard camera are processed on the Raspberry Pi to extract lane geometry and calculate steering angles,enabling the vehicle to follow the planned path.In addition,ultrasonic sensors placed in three directions at the front of the vehicle detect obstacles and allow real-time path adjustment for safe navigation.Experimental results demonstrate stable performance under controlled conditions,highlighting the system’s potential for scalable autonomous driving applications.This work confirms that deep learning methods can be efficiently deployed on low-power embedded systems,offering a practical framework for navigation,path planning,and intelligent transportation research.展开更多
In the present work,autonomous mobile robot(AMR)system is intended with basic behaviour,one is obstacle avoidance and the other is target seeking in various environments.The AMR is navigated using fuzzy logic,neural n...In the present work,autonomous mobile robot(AMR)system is intended with basic behaviour,one is obstacle avoidance and the other is target seeking in various environments.The AMR is navigated using fuzzy logic,neural network and adaptive neurofuzzy inference system(ANFIS)controller with safe boundary algorithm.In this method of target seeking behaviour,the obstacle avoidance at every instant improves the performance of robot in navigation approach.The inputs to the controller are the signals from various sensors fixed at front face,left and right face of the AMR.The output signal from controller regulates the angular velocity of both front power wheels of the AMR.The shortest path is identified using fuzzy,neural network and ANFIS techniques with integrated safe boundary algorithm and the predicted results are validated with experimentation.The experimental result has proven that ANFIS with safe boundary algorithm yields better performance in navigation,in particular with curved/irregular obstacles.展开更多
An irmovative mobile robot that has reconfigurable loeomotion chassis and reconfigurable bionic wheels has been developed to meet the needs of different payload and different terrain. Several prototypes have been achi...An irmovative mobile robot that has reconfigurable loeomotion chassis and reconfigurable bionic wheels has been developed to meet the needs of different payload and different terrain. Several prototypes have been achieved by the recortfiguration. By modeling relative comparison coefficients, these prototypes are analyzed in terms of geometrical parameter of trafficability, static stability and maneuverability. The effects of reconfiguration on these indices of robot performance can be compared, i.e. the variable height of chassis h has the biggest effect, the variable length of chassis 1 is the second, then is the camber angle β and the caster angle α. Some principles for reconfiguration are proposed.展开更多
In the mobile robotic systems a precise estimate of the robot pose (Cartesian [x y] position plus orientation angle theta) with the intention of the path planning optimization is essential for the correct performance,...In the mobile robotic systems a precise estimate of the robot pose (Cartesian [x y] position plus orientation angle theta) with the intention of the path planning optimization is essential for the correct performance, on the part of the robots, for tasks that are destined to it, especially when intention is for mobile robot autonomous navigation. This work uses a ToF (Time-of-Flight) of the RF digital signal interacting with beacons for computational triangulation in the way to provide a pose estimative at bi-dimensional indoor environment, where GPS system is out of range. It’s a new technology utilization making good use of old ultrasonic ToF methodology that takes advantage of high performance multicore DSP processors to calculate ToF of the order about ns. Sensors data like odometry, compass and the result of triangulation Cartesian estimative, are fused in a Kalman filter in the way to perform optimal estimation and correct robot pose. A mobile robot platform with differential drive and nonholonomic constraints is used as base for state space, plants and measurements models that are used in the simulations and for validation the experiments.展开更多
Precise and robust three-dimensional object detection(3DOD)presents a promising opportunity in the field of mobile robot(MR)navigation.Monocular 3DOD techniques typically involve extending existing twodimensional obje...Precise and robust three-dimensional object detection(3DOD)presents a promising opportunity in the field of mobile robot(MR)navigation.Monocular 3DOD techniques typically involve extending existing twodimensional object detection(2DOD)frameworks to predict the three-dimensional bounding box(3DBB)of objects captured in 2D RGB images.However,these methods often require multiple images,making them less feasible for various real-time scenarios.To address these challenges,the emergence of agile convolutional neural networks(CNNs)capable of inferring depth froma single image opens a new avenue for investigation.The paper proposes a novel ELDENet network designed to produce cost-effective 3DBounding Box Estimation(3D-BBE)froma single image.This novel framework comprises the PP-LCNet as the encoder and a fast convolutional decoder.Additionally,this integration includes a Squeeze-Exploit(SE)module utilizing the Math Kernel Library for Deep Neural Networks(MKLDNN)optimizer to enhance convolutional efficiency and streamline model size during effective training.Meanwhile,the proposed multi-scale sub-pixel decoder generates high-quality depth maps while maintaining a compact structure.Furthermore,the generated depthmaps provide a clear perspective with distance details of objects in the environment.These depth insights are combined with 2DOD for precise evaluation of 3D Bounding Boxes(3DBB),facilitating scene understanding and optimal route planning for mobile robots.Based on the estimated object center of the 3DBB,the Deep Reinforcement Learning(DRL)-based obstacle avoidance strategy for MRs is developed.Experimental results demonstrate that our model achieves state-of-the-art performance across three datasets:NYU-V2,KITTI,and Cityscapes.Overall,this framework shows significant potential for adaptation in intelligent mechatronic systems,particularly in developing knowledge-driven systems for mobile robot navigation.展开更多
It is discussed with the design and implementation of an architecture for a mobile robot to navigate in dynamic and anknown indoor environments. The architecture is based on the framework of Open Robot Control Softwar...It is discussed with the design and implementation of an architecture for a mobile robot to navigate in dynamic and anknown indoor environments. The architecture is based on the framework of Open Robot Control Software at KTH （OROCOS@KTH）, which is also discussed and evaluated to navigate indoor efficiently, a new algorithm named door-like-exit detection is proposed which employs 2D feature oft. door and extracts key points of pathway from the raw data of a laser scanner. As a hybrid architecture, it is decomposed into several basic components which can be classified as either deliberative or reactive. Each component can concurrently execute and communicate with another. It is expansible and transferable and its components are reusable.展开更多
Due to the necessity for lightweight and efficient network models, deploying semantic segmentation models on mobile robots (MRs) is a formidable task. The fundamental limitation of the problem lies in the training per...Due to the necessity for lightweight and efficient network models, deploying semantic segmentation models on mobile robots (MRs) is a formidable task. The fundamental limitation of the problem lies in the training performance, the ability to effectively exploit the dataset, and the ability to adapt to complex environments when deploying the model. By utilizing the knowledge distillation techniques, the article strives to overcome the above challenges with the inheritance of the advantages of both the teacher model and the student model. More precisely, the ResNet152-PSP-Net model’s characteristics are utilized to train the ResNet18-PSP-Net model. Pyramid pooling blocks are utilized to decode multi-scale feature maps, creating a complete semantic map inference. The student model not only preserves the strong segmentation performance from the teacher model but also improves the inference speed of the prediction results. The proposed method exhibits a clear advantage over conventional convolutional neural network (CNN) models, as evident from the conducted experiments. Furthermore, the proposed model also shows remarkable improvement in processing speed when compared with light-weight models such as MobileNetV2 and EfficientNet based on latency and throughput parameters. The proposed KD-SegNet model obtains an accuracy of 96.3% and a mIoU (mean Intersection over Union) of 77%, outperforming the performance of existing models by more than 15% on the same training dataset. The suggested method has an average training time that is only 0.51 times less than same field models, while still achieving comparable segmentation performance. Hence, the semantic segmentation frames are collected, forming the motion trajectory for the system in the environment. Overall, this architecture shows great promise for the development of knowledge-based systems for MR’s navigation.展开更多
Navigation is a fundamental problem of mobile robots,for which Deep Reinforcement Learning(DRL)has received significant attention because of its strong representation and experience learning abilities.There is a growi...Navigation is a fundamental problem of mobile robots,for which Deep Reinforcement Learning(DRL)has received significant attention because of its strong representation and experience learning abilities.There is a growing trend of applying DRL to mobile robot navigation.In this paper,we review DRL methods and DRL-based navigation frameworks.Then we systematically compare and analyze the relationship and differences between four typical application scenarios:local obstacle avoidance,indoor navigation,multi-robot navigation,and social navigation.Next,we describe the development of DRL-based navigation.Last,we discuss the challenges and some possible solutions regarding DRL-based navigation.展开更多
A novel method was designed to solve reinforcement learning problems with artificial potential field.Firstly a reinforcement learning problem was transferred to a path planning problem by using artificial potential fi...A novel method was designed to solve reinforcement learning problems with artificial potential field.Firstly a reinforcement learning problem was transferred to a path planning problem by using artificial potential field(APF),which was a very appropriate method to model a reinforcement learning problem.Secondly,a new APF algorithm was proposed to overcome the local minimum problem in the potential field methods with a virtual water-flow concept.The performance of this new method was tested by a gridworld problem named as key and door maze.The experimental results show that within 45 trials,good and deterministic policies are found in almost all simulations.In comparison with WIERING's HQ-learning system which needs 20 000 trials for stable solution,the proposed new method can obtain optimal and stable policy far more quickly than HQ-learning.Therefore,the new method is simple and effective to give an optimal solution to the reinforcement learning problem.展开更多
文摘Recent advancements in autonomous vehicle technologies are transforming intelligent transportation systems.Artificial intelligence enables real-time sensing,decision-making,and control on embedded platforms with improved efficiency.This study presents the design and implementation of an autonomous radio-controlled(RC)vehicle prototype capable of lane line detection,obstacle avoidance,and navigation through dynamic path planning.The system integrates image processing and ultrasonic sensing,utilizing Raspberry Pi for vision-based tasks and ArduinoNano for real-time control.Lane line detection is achieved through conventional image processing techniques,providing the basis for local path generation,while traffic sign classification employs a You Only Look Once(YOLO)model optimized with TensorFlow Lite to support navigation decisions.Images captured by the onboard camera are processed on the Raspberry Pi to extract lane geometry and calculate steering angles,enabling the vehicle to follow the planned path.In addition,ultrasonic sensors placed in three directions at the front of the vehicle detect obstacles and allow real-time path adjustment for safe navigation.Experimental results demonstrate stable performance under controlled conditions,highlighting the system’s potential for scalable autonomous driving applications.This work confirms that deep learning methods can be efficiently deployed on low-power embedded systems,offering a practical framework for navigation,path planning,and intelligent transportation research.
文摘In the present work,autonomous mobile robot(AMR)system is intended with basic behaviour,one is obstacle avoidance and the other is target seeking in various environments.The AMR is navigated using fuzzy logic,neural network and adaptive neurofuzzy inference system(ANFIS)controller with safe boundary algorithm.In this method of target seeking behaviour,the obstacle avoidance at every instant improves the performance of robot in navigation approach.The inputs to the controller are the signals from various sensors fixed at front face,left and right face of the AMR.The output signal from controller regulates the angular velocity of both front power wheels of the AMR.The shortest path is identified using fuzzy,neural network and ANFIS techniques with integrated safe boundary algorithm and the predicted results are validated with experimentation.The experimental result has proven that ANFIS with safe boundary algorithm yields better performance in navigation,in particular with curved/irregular obstacles.
文摘An irmovative mobile robot that has reconfigurable loeomotion chassis and reconfigurable bionic wheels has been developed to meet the needs of different payload and different terrain. Several prototypes have been achieved by the recortfiguration. By modeling relative comparison coefficients, these prototypes are analyzed in terms of geometrical parameter of trafficability, static stability and maneuverability. The effects of reconfiguration on these indices of robot performance can be compared, i.e. the variable height of chassis h has the biggest effect, the variable length of chassis 1 is the second, then is the camber angle β and the caster angle α. Some principles for reconfiguration are proposed.
文摘In the mobile robotic systems a precise estimate of the robot pose (Cartesian [x y] position plus orientation angle theta) with the intention of the path planning optimization is essential for the correct performance, on the part of the robots, for tasks that are destined to it, especially when intention is for mobile robot autonomous navigation. This work uses a ToF (Time-of-Flight) of the RF digital signal interacting with beacons for computational triangulation in the way to provide a pose estimative at bi-dimensional indoor environment, where GPS system is out of range. It’s a new technology utilization making good use of old ultrasonic ToF methodology that takes advantage of high performance multicore DSP processors to calculate ToF of the order about ns. Sensors data like odometry, compass and the result of triangulation Cartesian estimative, are fused in a Kalman filter in the way to perform optimal estimation and correct robot pose. A mobile robot platform with differential drive and nonholonomic constraints is used as base for state space, plants and measurements models that are used in the simulations and for validation the experiments.
文摘Precise and robust three-dimensional object detection(3DOD)presents a promising opportunity in the field of mobile robot(MR)navigation.Monocular 3DOD techniques typically involve extending existing twodimensional object detection(2DOD)frameworks to predict the three-dimensional bounding box(3DBB)of objects captured in 2D RGB images.However,these methods often require multiple images,making them less feasible for various real-time scenarios.To address these challenges,the emergence of agile convolutional neural networks(CNNs)capable of inferring depth froma single image opens a new avenue for investigation.The paper proposes a novel ELDENet network designed to produce cost-effective 3DBounding Box Estimation(3D-BBE)froma single image.This novel framework comprises the PP-LCNet as the encoder and a fast convolutional decoder.Additionally,this integration includes a Squeeze-Exploit(SE)module utilizing the Math Kernel Library for Deep Neural Networks(MKLDNN)optimizer to enhance convolutional efficiency and streamline model size during effective training.Meanwhile,the proposed multi-scale sub-pixel decoder generates high-quality depth maps while maintaining a compact structure.Furthermore,the generated depthmaps provide a clear perspective with distance details of objects in the environment.These depth insights are combined with 2DOD for precise evaluation of 3D Bounding Boxes(3DBB),facilitating scene understanding and optimal route planning for mobile robots.Based on the estimated object center of the 3DBB,the Deep Reinforcement Learning(DRL)-based obstacle avoidance strategy for MRs is developed.Experimental results demonstrate that our model achieves state-of-the-art performance across three datasets:NYU-V2,KITTI,and Cityscapes.Overall,this framework shows significant potential for adaptation in intelligent mechatronic systems,particularly in developing knowledge-driven systems for mobile robot navigation.
基金The project is supported by European Open Robot Control Software Founda-tion(No.IST-2000-31064), National Natural Science Foundation of China(No.60475031) and the Swedish Foundation for Strategic Research, Sweden.
文摘It is discussed with the design and implementation of an architecture for a mobile robot to navigate in dynamic and anknown indoor environments. The architecture is based on the framework of Open Robot Control Software at KTH （OROCOS@KTH）, which is also discussed and evaluated to navigate indoor efficiently, a new algorithm named door-like-exit detection is proposed which employs 2D feature oft. door and extracts key points of pathway from the raw data of a laser scanner. As a hybrid architecture, it is decomposed into several basic components which can be classified as either deliberative or reactive. Each component can concurrently execute and communicate with another. It is expansible and transferable and its components are reusable.
基金funded by Hanoi University of Science and Technology(HUST)under project number T2023-PC-008.
文摘Due to the necessity for lightweight and efficient network models, deploying semantic segmentation models on mobile robots (MRs) is a formidable task. The fundamental limitation of the problem lies in the training performance, the ability to effectively exploit the dataset, and the ability to adapt to complex environments when deploying the model. By utilizing the knowledge distillation techniques, the article strives to overcome the above challenges with the inheritance of the advantages of both the teacher model and the student model. More precisely, the ResNet152-PSP-Net model’s characteristics are utilized to train the ResNet18-PSP-Net model. Pyramid pooling blocks are utilized to decode multi-scale feature maps, creating a complete semantic map inference. The student model not only preserves the strong segmentation performance from the teacher model but also improves the inference speed of the prediction results. The proposed method exhibits a clear advantage over conventional convolutional neural network (CNN) models, as evident from the conducted experiments. Furthermore, the proposed model also shows remarkable improvement in processing speed when compared with light-weight models such as MobileNetV2 and EfficientNet based on latency and throughput parameters. The proposed KD-SegNet model obtains an accuracy of 96.3% and a mIoU (mean Intersection over Union) of 77%, outperforming the performance of existing models by more than 15% on the same training dataset. The suggested method has an average training time that is only 0.51 times less than same field models, while still achieving comparable segmentation performance. Hence, the semantic segmentation frames are collected, forming the motion trajectory for the system in the environment. Overall, this architecture shows great promise for the development of knowledge-based systems for MR’s navigation.
文摘Navigation is a fundamental problem of mobile robots,for which Deep Reinforcement Learning(DRL)has received significant attention because of its strong representation and experience learning abilities.There is a growing trend of applying DRL to mobile robot navigation.In this paper,we review DRL methods and DRL-based navigation frameworks.Then we systematically compare and analyze the relationship and differences between four typical application scenarios:local obstacle avoidance,indoor navigation,multi-robot navigation,and social navigation.Next,we describe the development of DRL-based navigation.Last,we discuss the challenges and some possible solutions regarding DRL-based navigation.
基金Projects(30270496,60075019,60575012)supported by the National Natural Science Foundation of China
文摘A novel method was designed to solve reinforcement learning problems with artificial potential field.Firstly a reinforcement learning problem was transferred to a path planning problem by using artificial potential field(APF),which was a very appropriate method to model a reinforcement learning problem.Secondly,a new APF algorithm was proposed to overcome the local minimum problem in the potential field methods with a virtual water-flow concept.The performance of this new method was tested by a gridworld problem named as key and door maze.The experimental results show that within 45 trials,good and deterministic policies are found in almost all simulations.In comparison with WIERING's HQ-learning system which needs 20 000 trials for stable solution,the proposed new method can obtain optimal and stable policy far more quickly than HQ-learning.Therefore,the new method is simple and effective to give an optimal solution to the reinforcement learning problem.
