Towards Real-Time Path Planning through Deep Reinforcement Learning for a UAV in Dynamic Environments

Yan, Chao; Xiang, Xiaojia; Wang, Chang

doi:10.1007/s10846-019-01073-3

Cited by 171 publications

(55 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The replay buffer size is 10 6 . The above hyperparameter settings are referenced from [42]. In addition, the exploration rate ε linearly decreased from 0.7 to 0.1 over a period of 1 million steps then fixed.…”

Section: Network Architecture and Hyperparameter Settingmentioning

confidence: 99%

Energy-Optimal Flight Strategy for Solar-Powered Aircraft Using Reinforcement Learning With Discrete Actions

2021

IEEE Access

View full text Add to dashboard Cite

The low efficiency of photovoltaic cells limits the energy absorption of high-altitude longendurance (HALE) solar-powered unmanned aircraft vehicles (UAVs), which dramatically weakens the capacity for long-endurance missions. Therefore, finding a method to extend the flight duration with finite solar energy drives extensive research. The present work introduces a method that applies a deep reinforcement learning (DRL) framework to generate an energy-optimized flight strategy for HALE solarpowered aircraft. The neural network controller is designed to realize autonomous flight navigation by giving commands of thrust, attack angle, and bank angle. A mission area with a radius of 5 km is assumed to test the RL controller performance. The simulation results show that the RL controller leads to a 28 % increase in the battery SoC after a 24-hour flight, which indicates that a controller based on the RL framework might be a potential method for solving the solar-powered UAV trajectory planning problem. Aiming to explore the applicability of the RL controller, a sustained flight test is implemented. The results show that a 39-day endurance flight is achieved by the RL controller, which is 50% higher than the base case with a steady flight trajectory.

show abstract

Section: Network Architecture and Hyperparameter Settingmentioning

confidence: 99%

Energy-Optimal Flight Strategy for Solar-Powered Aircraft Using Reinforcement Learning With Discrete Actions

2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…As the pivotal foundation in social robot navigation, a wide variety of path planning methods have been proposed for robots navigating in different environments [7][8][9][10]. In general, these methods can be divided into global and local methods according to the completeness of map information known before path planning progress.…”

Section: Related Workmentioning

confidence: 99%

A Dynamic Path Planning Method for Social Robots in the Home Environment

Zhang

Zhou

et al. 2020

Electronics

View full text Add to dashboard Cite

The home environment is a typical dynamic environment with moveable obstacles. The social robots working in home need to search for feasible paths in this complex dynamic environment. In this work, we propose an improved RRT algorithm to plan feasible path in home environment. The algorithm pre-builds a tree that covers the whole map and maintains the effectiveness of all nodes with branch pruning, reconnection, and regrowth process. The method forms a path by searching the nearest node in the tree and then quickly accessing the nodes near the destination. Due to the effectiveness-maintaining process, the proposed method can effectively deal with the complex dynamic environment where the destination and multiple moving obstacles change simultaneously. In addition, our method can be extended to the path-planning problem in 3D space. The simulation experiments verify the effectiveness of the algorithm.

show abstract

“…More critical situation such as scenarios with disaster and dynamic threats usually desires improved intelligent algorithms [18][19][20]. More recently, the development of deep learning also spawned the deep reinforcement learning based path planning [24].…”

Section: Introductionmentioning

confidence: 99%

Path Planning for Unmanned Aerial Vehicle Using a Mix-Strategy-Based Gravitational Search Algorithm

Jiang

Zhang

2021

IEEE Access

View full text Add to dashboard Cite

Path planning is a global optimization problem aims to program the optimal flight path for Unmanned Aerial Vehicle (UAV) that has short length and suffers from low threat. In this paper, we present a Mixed-Strategy based Gravitational Search Algorithm (MSGSA) for the path planning. In MSGSA, an adaptive adjustment strategy for the gravitational constant attenuation factor alpha (α) is presented firstly, in which the value of α is adjusted based on the evolutionary state of the particles. This helps to adaptively balance the exploration and exploitation of the algorithm. In addition, to further alleviate the premature convergence problem, a Cauchy mutation strategy is developed for MSGSA. In this strategy, only when the global best particle cannot be further improved for several times the mutation is executed. In the MSGSA based path planning procedure, we construct an objective function using the flight length cost, threat area cost, and turning angle constraint to decrease the flight risk and obtain the smoother path. For performance evaluation, the MSGSA is applied to two typical simulated flight missions with complex flight environments, including user-defined forbidden flying areas, Radar, missile, artillery and anti-aircraft gun. The obtained flight paths are compared with that of the standard Gravitational Search Algorithm (GSA) and two improved variants of GSA, i.e. gbest-guided GSA (GGSA), and hybrid Particle Swarm Optimization and GSA (PSOGSA). The experimental results demonstrate the superiority of the MSGSA based method in terms of the solution quality, robustness, as well as the constraint-handling ability.

show abstract

Towards Real-Time Path Planning through Deep Reinforcement Learning for a UAV in Dynamic Environments

Cited by 171 publications

References 9 publications

Energy-Optimal Flight Strategy for Solar-Powered Aircraft Using Reinforcement Learning With Discrete Actions

Energy-Optimal Flight Strategy for Solar-Powered Aircraft Using Reinforcement Learning With Discrete Actions

A Dynamic Path Planning Method for Social Robots in the Home Environment

Path Planning for Unmanned Aerial Vehicle Using a Mix-Strategy-Based Gravitational Search Algorithm

Contact Info

Product

Resources

About