Neural RRT*: Learning-Based Optimal Path Planning

Wang, Jiankun; Chi, Wenzheng; Li, Chenming; Wang, Chaoqun; Meng, Max Q.-H.

doi:10.1109/tase.2020.2976560

Cited by 295 publications

(142 citation statements)

References 27 publications

Supporting

Mentioning

141

Contrasting

Unclassified

Order By: Relevance

“…The Neural RRT* algorithm, proposed by Wang et al in Reference [ 21 ], is a novel optimal path planning algorithm based on convolutional neural networks. It used the A* algorithm to generate training data, considering map information as input, and the optimal path as ground truth.…”

Section: Resultsmentioning

confidence: 99%

“…Novel DNN-based path planning methods which have been developed are based on biologically inspired cognitive architectures [ 20 ], with the three primary methods being swarm intelligence, evolutionary algorithms, and neurodynamics. In the case of Reference [ 21 ], an optimal path planning algorithm based on convolutional neural networks (CNN) and random-exporing trees (RRT) is presented. Their approach, called Neural RRT*, is a framework for generating the sampling distribution of the optimal path under several constraints.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

OctoPath: An OcTree-Based Self-Supervised Learning Approach to Local Trajectory Planning for Mobile Robots

Trasnea

Ginerica

Zaha

et al. 2021

Sensors

View full text Add to dashboard Cite

Autonomous mobile robots are usually faced with challenging situations when driving in complex environments. Namely, they have to recognize the static and dynamic obstacles, plan the driving path and execute their motion. For addressing the issue of perception and path planning, in this paper, we introduce OctoPath, which is an encoder-decoder deep neural network, trained in a self-supervised manner to predict the local optimal trajectory for the ego-vehicle. Using the discretization provided by a 3D octree environment model, our approach reformulates trajectory prediction as a classification problem with a configurable resolution. During training, OctoPath minimizes the error between the predicted and the manually driven trajectories in a given training dataset. This allows us to avoid the pitfall of regression-based trajectory estimation, in which there is an infinite state space for the output trajectory points. Environment sensing is performed using a 40-channel mechanical LiDAR sensor, fused with an inertial measurement unit and wheels odometry for state estimation. The experiments are performed both in simulation and real-life, using our own developed GridSim simulator and RovisLab’s Autonomous Mobile Test Unit platform. We evaluate the predictions of OctoPath in different driving scenarios, both indoor and outdoor, while benchmarking our system against a baseline hybrid A-Star algorithm and a regression-based supervised learning method, as well as against a CNN learning-based optimal path planning method.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

OctoPath: An OcTree-Based Self-Supervised Learning Approach to Local Trajectory Planning for Mobile Robots

Trasnea

Ginerica

Zaha

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…The above aims to enhance the path-planning effectiveness by decreasing the implicit drawbacks of a single category. An example of these techniques is found in [21]. In the above research, an NN improves the RRT* algorithm by determining a suitable sampling distribution.…”

Section: Introduction 1a Review Of Path-planning Methodsmentioning

confidence: 99%

Path-Planning for Mobile Robots Using a Novel Variable-Length Differential Evolution Variant

et al. 2021

View full text Add to dashboard Cite

Mobile robots are currently exploited in various applications to enhance efficiency and reduce risks in hard activities for humans. The high autonomy in those systems is strongly related to the path-planning task. The path-planning problem is complex and requires in its formulation the adjustment of path elements that take the mobile robot from a start point to a target one at the lowest cost. Nevertheless, the identity or the number of the path elements to be adjusted is unknown; therefore, the human decision is necessary to determine this information reducing autonomy. Due to the above, this work conceives the path-planning as a Variable-Length-Vector optimization problem (VLV-OP) where both the number of variables (path elements) and their values must be determined. For this, a novel variant of Differential Evolution for Variable-Length-Vector optimization named VLV-DE is proposed to handle the path-planning VLV-OP for mobile robots. VLV-DE uses a population with solution vectors of different sizes adapted through a normalization procedure to allow interactions and determine the alternatives that better fit the problem. The effectiveness of this proposal is shown through the solution of the path-planning problem in complex scenarios. The results are contrasted with the well-known A* and the RRT*-Smart path-planning methods.

show abstract

“…Therefore, the adaptive mechanism of the proposed algorithm can show better performance when the problem is more complicated. The adaptive update method is determined by formulas (11) and (12).…”

Section: Fitness Evaluationmentioning

confidence: 99%

“…At this stage, there are many path planning algorithms that can be used for automated robot inspections of IWSNs, such as rapidly exploring random tree (RRT) algorithm, Dijkstra algorithm, A * algorithm, GA, and ICA [11][12][13][14][15]. These algorithms can find a relatively good path, but they also have various disadvantages.…”

Section: Introductionmentioning

confidence: 99%

An Adaptive Immune Ant Colony Optimization for Reducing Energy Consumption of Automatic Inspection Path Planning in Industrial Wireless Sensor Networks

Xiao

Liu

et al. 2021

Journal of Sensors

View full text Add to dashboard Cite

Industrial wireless sensor networks (IWSNs) are usually fixedly deployed in industrial environments, and various sensor nodes cooperate with each other to complete industrial production tasks. The efficient work of each sensor node of IWSNs will improve the efficiency of the entire network. Automated robots need to perform timely inspection and maintenance of IWSNs in an industrial environment. Excessive inspection distance will increase inspection costs and increase energy consumption. Therefore, shortening the inspection distance can reduce production energy consumption, which is very important for the efficient operation of the entire system. However, the optimal detection path planning of IWSNs is an N-P problem, which can usually only be solved by heuristic mathematical methods. This paper proposes a new adaptive immune ant colony optimization (AIACO) for optimizing automated inspection path planning. Moreover, novel adaptive operator and immune operator are designed to prevent the algorithm from falling into the local optimum and increase the optimization ability. In order to verify the performance of the algorithm, the algorithm is compared with genetic algorithm (GA) and immune clone algorithm (ICA). The simulation results show that the inspection distance of IWSNs using AIACO is lower than that of GA and ICA. In addition, the convergence speed of AIACO is faster than that of GA and ICA. Therefore, the AIACO proposed in this paper can effectively reduce the inspection energy consumption of the entire IWSN system.

show abstract

Neural RRT*: Learning-Based Optimal Path Planning

Cited by 295 publications

References 27 publications

OctoPath: An OcTree-Based Self-Supervised Learning Approach to Local Trajectory Planning for Mobile Robots

OctoPath: An OcTree-Based Self-Supervised Learning Approach to Local Trajectory Planning for Mobile Robots

Path-Planning for Mobile Robots Using a Novel Variable-Length Differential Evolution Variant

An Adaptive Immune Ant Colony Optimization for Reducing Energy Consumption of Automatic Inspection Path Planning in Industrial Wireless Sensor Networks

Contact Info

Product

Resources

About