Showing papers on "Collision avoidance published in 2019"
TL;DR: A decentralized adaptive formation controller is designed that ensures uniformly ultimate boundedness of the closed-loop system with prescribed performance and avoids collision between each vehicle and its leader.
Abstract: This paper addresses a decentralized leader–follower formation control problem for a group of fully actuated unmanned surface vehicles with prescribed performance and collision avoidance. The vehicles are subject to time-varying external disturbances, and the vehicle dynamics include both parametric uncertainties and uncertain nonlinear functions. The control objective is to make each vehicle follow its reference trajectory and avoid collision between each vehicle and its leader. We consider prescribed performance constraints, including transient and steady-state performance constraints, on formation tracking errors. In the kinematic design, we introduce the dynamic surface control technique to avoid the use of vehicle's acceleration. To compensate for the uncertainties and disturbances, we apply an adaptive control technique to estimate the uncertain parameters including the upper bounds of the disturbances and present neural network approximators to estimate uncertain nonlinear dynamics. Consequently, we design a decentralized adaptive formation controller that ensures uniformly ultimate boundedness of the closed-loop system with prescribed performance and avoids collision between each vehicle and its leader. Simulation results illustrate the effectiveness of the decentralized formation controller.
273 citations
TL;DR: A modified Artificial Potential Field (APF), which contains a new modified repulsion potential field function and the corresponding virtual forces, is developed to address the issue of Collision Avoidance with dynamic targets and static obstacles, including emergency situations.
Abstract: This paper presents a real-time and deterministic path planning method for autonomous ships or Unmanned Surface Vehicles (USV) in complex and dynamic navigation environments. A modified Artificial Potential Field (APF), which contains a new modified repulsion potential field function and the corresponding virtual forces, is developed to address the issue of Collision Avoidance (CA) with dynamic targets and static obstacles, including emergency situations. Appropriate functional and safety requirements are added in the corresponding virtual forces to ensure International Regulations for Preventing Collisions at Sea (COLREGS)-constrained behaviour for the own ship's CA actions. Simulations show that the method is fast, effective and deterministic for path planning in complex situations with multiple moving target ships and stationary obstacles and can account for the unpredictable strategies of other ships. The authors believe that automatic navigation systems operated without human interaction could benefit from the development of path planning algorithms.
184 citations
16 Jan 2019
TL;DR: A tight bound for approximation of collision probability is developed, which makes the CCNMPC formulation tractable and solvable in real time.
Abstract: Safe autonomous navigation of microair vehicles in cluttered dynamic environments is challenging due to the uncertainties arising from robot localization, sensing, and motion disturbances. This letter presents a probabilistic collision avoidance method for navigation among other robots and moving obstacles, such as humans. The approach explicitly considers the collision probability between each robot and obstacle and formulates a chance constrained nonlinear model predictive control problem (CCNMPC). A tight bound for approximation of collision probability is developed, which makes the CCNMPC formulation tractable and solvable in real time. For multirobot coordination, we describe three approaches, one distributed without communication (constant velocity assumption), one distributed with communication (of previous plans), and one centralized (sequential planning). We evaluate the proposed method in experiments with two quadrotors sharing the space with two humans and verify the multirobot coordination strategy in simulation with up to sixteen quadrotors.
179 citations
TL;DR: This study proposes an efficient method to overcome multiship collision avoidance problems based on the Deep Reinforcement Learning (DRL) algorithm and demonstrates its excellent adaptability to unknown complex environments with various encountered ships.
127 citations
TL;DR: The proposed GVO-CAS can offer rule-compliant evasive actions with a minimum number of required actions for ships and shows the great potential to use the GVO algorithm in both manned and unmanned ships at sea.
124 citations
TL;DR: A novel approach based on deep reinforcement learning (DRL) is proposed for automatic collision avoidance of multiple ships particularly in restricted waters, incorporating ship manoeuvrability, human experience and navigation rules.
121 citations
TL;DR: This paper presents a task analysis for collision avoidance through Hierarchical Task Analysis and making use of a cognitive model for categorizing the tasks, and identifies human failure events in future MASS operations.
119 citations
TL;DR: A deep neural network is used to approximate the table, reducing the required storage space by a factor of 1000 and enabling the collision avoidance system to operate using current avionics systems.
Abstract: One approach to designing decision-making logic for an aircraft collision avoidance system frames the problem as a Markov decision process and optimizes the system using dynamic programming. The re...
116 citations
TL;DR: This work aims to provide researchers with a state-of-the-art overview of various approaches for multi-UAV collision avoidance through several classifications based on algorithm used and frameworks designed.
109 citations
TL;DR: This article reviews the most recent developments on DL Unmanned Aerial Systems (UASs) and provides a detailed explanation on the main DL techniques, and enumerates the most relevant open challenges for current DL-UAV solutions, thus allowing future researchers to define a roadmap for devising the new generation affordable autonomous DL- UAV IoT solutions.
Abstract: Advances in Unmanned Aerial Vehicles (UAVs), also known as drones, offer unprecedented opportunities to boost a wide array of large-scale Internet of Things (IoT) applications. Nevertheless, UAV platforms still face important limitations mainly related to autonomy and weight that impact their remote sensing capabilities when capturing and processing the data required for developing autonomous and robust real-time obstacle detection and avoidance systems. In this regard, Deep Learning (DL) techniques have arisen as a promising alternative for improving real-time obstacle detection and collision avoidance for highly autonomous UAVs. This article reviews the most recent developments on DL Unmanned Aerial Systems (UASs) and provides a detailed explanation on the main DL techniques. Moreover, the latest DL-UAV communication architectures are studied and their most common hardware is analyzed. Furthermore, this article enumerates the most relevant open challenges for current DL-UAV solutions, thus allowing future researchers to define a roadmap for devising the new generation affordable autonomous DL-UAV IoT solutions.
96 citations
TL;DR: Together with artificial potential field (APF) method, an adaptive leader-following formation control with collision avoidance strategy is developed for a class of second-order nonlinear multi-agent systems that can achieve an ideal formation pattern with the collision avoidance performance.
TL;DR: In this article, Collision avoidance and stabilisation are two of the most crucial concerns when an autonomous vehicle finds itself in emergency situations, which usually occur in a short time horizon and require l...
Abstract: Collision avoidance and stabilisation are two of the most crucial concerns when an autonomous vehicle finds itself in emergency situations, which usually occur in a short time horizon and require l...
TL;DR: In this article, a distributed model predictive control (DMPC) algorithm was proposed to generate trajectories in real-time for multiple robots in point-to-point transition tasks.
Abstract: We present a distributed model predictive control (DMPC) algorithm to generate trajectories in real-time for multiple robots. We adopted the \textit{on-demand collision avoidance} method presented in previous work to efficiently compute non-colliding trajectories in transition tasks. An event-triggered replanning strategy is proposed to account for disturbances. Our simulation results show that the proposed collision avoidance method can reduce, on average, around 50% of the travel time required to complete a multi-agent point-to-point transition when compared to the well-studied Buffered Voronoi Cells (BVC) approach. Additionally, it shows a higher success rate in transition tasks with a high density of agents, with more than 90% success rate with 30 palm-sized quadrotor agents in a 18 m^3 arena. The approach was experimentally validated with a swarm of up to 20 drones flying in close proximity.
TL;DR: A sensor-based feedback law is constructed that provably solves the real-time collision-free robot navigation problem in a compact convex Euclidean subset cluttered with unknown but sufficiently separated and strongly convex obstacles and extended to practically motivated limited range, isotropic and anisotropic sensing models, and the standard differential-drive vehicle.
Abstract: We construct a sensor-based feedback law that provably solves the real-time collision-free robot navigation problem in a compact convex Euclidean subset cluttered with unknown but sufficiently sepa...
TL;DR: The superiority of the proposed swarm control strategy is that the distributed controller and flexible formation enable USVs to follow the target and avoid collisions autonomously.
Posted Content•
TL;DR: It is demonstrated that with the proposed framework, the simulator car agents can transfer knowledge to the RC cars in real-time, with 27% increase in the average distance with obstacles and 42% decrease in the collision counts.
Abstract: Reinforcement learning (RL) is widely used in autonomous driving tasks and training RL models typically involves in a multi-step process: pre-training RL models on simulators, uploading the pre-trained model to real-life robots, and fine-tuning the weight parameters on robot vehicles. This sequential process is extremely time-consuming and more importantly, knowledge from the fine-tuned model stays local and can not be re-used or leveraged collaboratively. To tackle this problem, we present an online federated RL transfer process for real-time knowledge extraction where all the participant agents make corresponding actions with the knowledge learned by others, even when they are acting in very different environments. To validate the effectiveness of the proposed approach, we constructed a real-life collision avoidance system with Microsoft Airsim simulator and NVIDIA JetsonTX2 car agents, which cooperatively learn from scratch to avoid collisions in indoor environment with obstacle objects. We demonstrate that with the proposed framework, the simulator car agents can transfer knowledge to the RC cars in real-time, with 27% increase in the average distance with obstacles and 42% decrease in the collision counts.
TL;DR: A distributed coordination strategy to deal with the many-to-many collision avoidance problem, which consists of two phases: predictions of ship trajectories are made based on ship dynamics, giving different candidate rudder angles, and potential collision risks that may be caused by each rudder angle selection are evaluated based on calculations of collision risk parameters.
TL;DR: This work proposes a MTS planner based on ant colony optimization that includes communication and collision avoidance constraints that ensures that the Unmanned Aerial Vehicles (UAVs) are able to complete the optimized search trajectories without risk of collision or loss of communication with the ground control station.
TL;DR: A means to quantify and subsequently evaluate the otherwise subjective nature of COLREGS thus providing a path toward standardized evaluation and certification of protocol-constrained collision avoidance systems based on admiralty case law and on-water experience is presented.
Abstract: Collision avoidance protocols such as COLREGS are written primarily for human operators resulting in a rule set that is open to some interpretation, difficult to quantify, and challenging to evaluate. Increasing use of autonomous control of vehicles emphasizes the need to more uniformly establish entry and exit criteria for collision avoidance rules, adopt a means to quantitatively evaluate performance, and establish a “road test” for autonomous marine vehicle collision avoidance. This paper presents a means to quantify and subsequently evaluate the otherwise subjective nature of COLREGS thus providing a path toward standardized evaluation and certification of protocol-constrained collision avoidance systems based on admiralty case law and on-water experience. Notional algorithms are presented for evaluation of COLREGS collision avoidance rules to include overtaking, head-on, crossing, give-way, and stand-on rules as well as applicable entry criteria. These rules complement and enable an autonomous collision avoidance road test as a first iteration of algorithm certification prior to vessels operating in human-present environments. Additional COLREGS rules are discussed for future development. Both real-time and post-mission protocol evaluation tools are introduced. While the motivation of these techniques applies to improvement of autonomous marine collision avoidance, the concepts for protocol evaluation and certification extend naturally to human-operated vessels. Evaluation of protocols governing other physical domains may also benefit from adapting these techniques to their cases.
TL;DR: A four-dimensional coordinated path planning algorithm for multiple UAVs is proposed, in which time variable is taken into account for each UAV as well as collision free and obstacle avoidance, to overcome the defects of local optimal and slow convergence.
TL;DR: In this study, an obstacle avoidance controller based on nonlinear model predictive control is designed in autonomous vehicle navigation and can ensure real-time trajectory tracking and collision avoidance.
Abstract: In this study, an obstacle avoidance controller based on nonlinear model predictive control is designed in autonomous vehicle navigation. The reference trajectory is predefined using a sigmoid function in accordance with road conditions. When obstacles suddenly appear on a predefined trajectory, the reference trajectory should be adjusted dynamically. For dynamic obstacles, a moving trend function is constructed to predict the obstacle position variances in the predictive horizon. Furthermore, a risk index is constructed and introduced into the cost function to realize collision avoidance by combining the relative position relationship between vehicle and obstacles in the predictive horizon. Meanwhile, lateral acceleration constraint is also considered to ensure vehicle stability. Finally, trajectory dynamic planning and tracking are integrated into a single-level model predictive controller. Simulation tests reveal that the designed controller can ensure real-time trajectory tracking and collision avoidance.
TL;DR: This paper first analyzes the feasibility conditions for a general continuous-time trajectory planning problem and then proposes an analytical solution method for two important boundary trajectory problems and proposes a discrete-time model with a more general objective function and a certain sparsity requirement that helps parsimonious planned trajectories.
Abstract: One challenging problem about connected automated vehicles is to optimize vehicle trajectories considering realistic constraints (e.g. vehicle kinematic limits and collision avoidance) and objectives (e.g., travel time, fuel consumption). With respect to communication cost and implementation difficulty, parsimonious trajectory planning has attracted continuous interests. In this paper, we first analyze the feasibility conditions for a general continuous-time trajectory planning problem and then propose an analytical solution method for two important boundary trajectory problems. We further propose a discrete-time model with a more general objective function and a certain sparsity requirement that helps parsimonious planned trajectories. This sparsity requirement is implemented with a l1 norm regulatory term appended to the objective function. Numerical examples are conducted on several representative applications and show that the proposed design strategy is effective.
TL;DR: Numerical results show that the proposed model provides an accurate estimation for car-following collision risk with a relatively low complexity, taking into account the impacts of vehicle dynamics, driver reaction capacity, and external environment on rear-end collisions.
Abstract: This paper presents a probabilistic decision-making framework for rear-end collision avoidance systems, focusing on modeling the impact of major collision-causing factors on the occurrence of accidents. Decisions on when and how to assist drivers are made using a Bayesian network approach according to collision risk evaluation results, given a prior probabilistic knowledge. The structure of the Bayesian network model is learnt using a K2 algorithm with a practical dataset. To provide adequate response time for drivers, we also predict collision probability in the next monitoring interval using a Kalman filter model. The prediction accuracy is evaluated with different use cases and compared to real scenarios with the obtained dataset. To make our framework more relevant to practical applications, we also discuss the corresponding safety control strategies by classifying collision risk into high and low levels. In addition, the proposed model is evaluated through experiments including simulations and road tests. In the simulations, the algorithms are tested in different scenarios with various configurations of weather conditions, driver response capability, and vehicular dynamics. In order to demonstrate collision avoidance performance, the proposed model is also compared to existing schemes. In the road tests, the algorithm is embedded into an unmanned vehicle with predefined parameters to evaluate the impact of computational complexity on rear-end collision avoidance. Numerical results show that the proposed model provides an accurate estimation for car-following collision risk with a relatively low complexity, taking into account the impacts of vehicle dynamics, driver reaction capacity, and external environment on rear-end collisions.
TL;DR: This study introduces the subject of manipulator’s on-line collision avoidance into a real industrial application implementing typical sensors and a commonly used collaborative industrial manipulator, KUKA iiwa.
TL;DR: A collision avoidance/mitigation system (CAMS) is proposed to rapidly evaluate risks associated with all surrounding vehicles and to maneuver the vehicle into a safer region when faced with critically dangerous situations.
Abstract: Despite development efforts toward autonomous vehicle technologies, the number of collisions and driver interventions of autonomous vehicles tested in California seems to be reaching a plateau. One of the main reasons for this is the lack of defensive driving functionality; i.e. emergency collision avoidance when other human drivers make mistakes. In this paper, a collision avoidance/mitigation system (CAMS) is proposed to rapidly evaluate risks associated with all surrounding vehicles and to maneuver the vehicle into a safer region when faced with critically dangerous situations. First, a risk assessment module, namely, predictive occupancy map (POM), is proposed to compute potential risks associated with surrounding vehicles based on relative position, velocity, and acceleration. Then, the safest trajectory with the lowest risk level is selected among the 12 local trajectories through the POM. To ensure stable and successful collision avoidance of the ego-vehicle, the lateral and longitudinal acceleration profiles are planned while considering the vehicle stability limit. The performance of the proposed algorithm is validated based on side and rear-end collision scenarios, which are difficult to predict and to monitor. The simulation results show that the proposed CAMS via POM detect a collision risk 1.4 s before the crash, and avoids the collision successfully.
TL;DR: An improved pseudo-random proportional rule is proposed to select the ant state transition and the wolf pack allocation principle and the maximum-minimum ant system are used to update the global pheromone to avoid the search falling into local optimum.
Abstract: In order to solve the problem of insufficient search ability of the unmanned surface vehicle (USV) collision avoidance planning algorithm, this paper proposes an improved ant colony optimization algorithm (ACO). First, aiming at the static unknown environment, in order to improve the real-time performance of USV online planning, and considering the environmental characteristics of USV operation for improving ACO to search for the optimal path, a dynamic viewable method is proposed for the local environment model. Second, according to the known dynamic environment, based on the motion velocity model and International Regulations for Preventing Collisions at Sea (COLREGS), a reverse eccentric expansion method is designed to deal with the dynamic obstacles. Then, aiming at the problem that ACO has a slow convergence speed, an improved pseudo-random proportional rule is proposed to select the ant state transition. And the wolf pack allocation principle and the maximum-minimum ant system are used to update the global pheromone to avoid the search falling into local optimum. Finally, the convergence, real-time performance, and stability of the improved ACO are verified through the simulation experiment of USV collision avoidance in the static unknown and dynamic known environment.
TL;DR: The results indicate that the improved DRL algorithm could effectively improve the navigation safety and collision avoidance.
Abstract: This research focuses on the adaptive navigation of maritime autonomous surface ships (MASSs) in an uncertain environment. To achieve intelligent obstacle avoidance of MASSs in a port, an autonomous navigation decision-making model based on hierarchical deep reinforcement learning is proposed. The model is mainly composed of two layers: the scene division layer and an autonomous navigation decision-making layer. The scene division layer mainly quantifies the sub-scenarios according to the International Regulations for Preventing Collisions at Sea (COLREG). This research divides the navigational situation of a ship into entities and attributes based on the ontology model and Protege language. In the decision-making layer, we designed a deep Q-learning algorithm utilizing the environmental model, ship motion space, reward function, and search strategy to learn the environmental state in a quantized sub-scenario to train the navigation strategy. Finally, two sets of verification experiments of the deep reinforcement learning (DRL) and improved DRL algorithms were designed with Rizhao port as a study case. Moreover, the experimental data were analyzed in terms of the convergence trend, iterative path, and collision avoidance effect. The results indicate that the improved DRL algorithm could effectively improve the navigation safety and collision avoidance.
TL;DR: A predictive collision avoidance method based on an improved beetle antennae search (BAS) algorithm for underactuated surface vessels is proposed, and an improved BAS algorithm is proposed to enhance the optimization performance of the original BAS algorithm under the known constraints, which is applied to solve the predictive collisions avoidance problem.
TL;DR: A novel nonlinear formation error function using the relative distance and angles between robots is introduced and the stability of the total closed-loop control system is analyzed in the Lyapunov sense.
TL;DR: A novel method to use a neural network to approximate an inverse model based on decisions made with MPC for collision avoidance in multi-ship encounters is proposed based on model predictive control, an improved Q-learning beetle swarm antenna search algorithm and neural networks.