Robot Navigation System in Stochastic Environment Based on Reinforcement Learning on Lidar Data

Dudarenko, Dmitry; Ковалев, А.Д.; Tolstoy, Ivan; Vatamaniuk, Irina

doi:10.1007/978-981-13-9267-2_44

Cited by 3 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Благодаря этому повышается эффективность роботов, движущихся по данным траекториям. В дальнейшем разработанный алгоритм будет применяться в задачах планирования маршрутов на с/х территориях [28][29][30][31] и в модульной робототехнике [32,33].…”

Section: выводыunclassified

Smoothing the Curvature of Trajectory of Ground Robot in 3D Space

Zakharov¹,

Савельев²

2021

Proceedings of the SWSU

View full text Add to dashboard Cite

Purpose or research. Development of an algorithm for smoothing the trajectory of a ground robot over rough terrain, represented as a graph in three-dimensional space. Methods. This article presents the CSA (Curve Smoothing and Averaging) algorithm for smoothing local curves in the Oxy plane that make up a global curve, represented as a path on a connected graph in 3D space. The presented algorithm is based on the previously developed LRLHD-A * approach, which uses information about the vertices of the graph, their neighbors and the edges connecting them to select the area through which the smoothed curve will run. In order to avoid a broken curve at the output of the algorithm, a curve averaging method was developed, the idea of which is to shift the midpoints of local curves along the edges on which they are located. Results. An experimental comparison was made of the curvature of the trajectories obtained using the curve smoothing algorithm with curve averaging (CSA) and without it (CS). The method was carried out on a threedimensional map of the area, presented in the form of a graph with 100082 vertices. For the experiments, a sample of 10 pairs of random vertices was used, between which a path was built using the LRLHD-A * algorithm. The results of the experiments have shown that averaging the curve after smoothing reduces its curvature from 24 to 42%. Conclusion. Trajectories smoothed using the developed CSA algorithm have smoother curve bends at turns, compared to the algorithm taken as a basis. This allows the robots to move more smoothly and, as a consequence, reduce the consumption of the robot's battery.

show abstract

Section: выводыunclassified

Smoothing the Curvature of Trajectory of Ground Robot in 3D Space

Zakharov¹,

Савельев²

2021

Proceedings of the SWSU

View full text Add to dashboard Cite

show abstract

“…The cognition model of the robot is based on an ANN, which can be used to control a robot to navigate a moving crowd efficiently and safely [33][34][35]. It has been shown by Dudarenko et al that data from a laser scanner can be used for navigation using a stochastic approach [36].…”

Section: Introduction and Related Workmentioning

confidence: 99%

Crowded Environment Navigation with NEAT: Impact of Perception Resolution on Controller Optimization

Seriani

Marcini

Caruso

et al. 2021

J Intell Robot Syst

View full text Add to dashboard Cite

Crowd navigation with autonomous systems is a topic which has seen a rapid increase in interest recently. While it appears natural to humans, being able to reach a target can prove difficult or impossible to a mobile robot because of the safety issues related to collisions with people. In this work we propose an approach to control a robot in a crowded environment; the method employs an Artificial Neural Network (ANN) that is trained with the NeuroEvolution of Augmented Topologies (NEAT) method. Models for the kinematics, perception, and cognition of the robot are presented. In particular, perception is based on a raycasting model which is tailored on the ANN. An in-depth analysis of a number of parameters of the environment and the robot is performed and a comparative analysis is presented; finally, results of the performance of the controller trained with NEAT are compared to those of a human driver who takes over the controller itself. Results show that the intelligent controller is able to perform on par with the human, within the simulated environment.

show abstract

Setting Artificial Neural Network Hyperparameters for Mobile Platform Navigation

Dudarenko

Смирнов

2020

Proceedings of the SWSU

View full text Add to dashboard Cite

Purpose of reseach. The main purpose of this work is to increase the efficiency of a neural network model when navigating a mobile robotic platform in static and dynamically generated environments. Methods. To solve this problem, precise setting and optimization of neural network hyperparameters were proposed. In order to encourage agents to explore the environment, the reward system was adjusted to increase the reward when the distance from the agent to the target point was reduced, and the penalty increased when moving in the opposite direction to the end point and passing each subsequent scene. This distribution of rewards and penalties encourages agents to learn actively and helps to reduce the total number of scenes. In order to reduce the amount of data processed by a neural network, normalization of input vectors was introduced. The learning time of the neural network model was reduced due to the parallel training of agents and, consequently, increased experience as a result of the environmental research. Results. The proposed approach reduced the learning time by 30% and improved the navigation efficiency of the mobile platform by 10% in a dynamically generated environment and by 22% in a static environment compared to the non-optimized model. Conclusion. The proposed solution can be used in conjunction with other methods of tracing and navigation, when the taught neural network works simultaneously with the already developed and proven navigation algorithms, for example, if the mobile platform connects a taught neural network only to adjust the position in space and to prevent collisions with other objects.

show abstract

Robot Navigation System in Stochastic Environment Based on Reinforcement Learning on Lidar Data

Cited by 3 publications

References 17 publications

Smoothing the Curvature of Trajectory of Ground Robot in 3D Space

Smoothing the Curvature of Trajectory of Ground Robot in 3D Space

Crowded Environment Navigation with NEAT: Impact of Perception Resolution on Controller Optimization

Setting Artificial Neural Network Hyperparameters for Mobile Platform Navigation

Contact Info

Product

Resources

About