Trajectory Planning for a Mobile Robot in a Dynamic Environment Using an LSTM Neural Network

Abstract

Autonomous mobile robots are an important focus of current research due to the advantages they bring to the industry, such as performing dangerous tasks with greater precision than humans. An autonomous mobile robot must be able to generate a collision-free trajectory while avoiding static and dynamic obstacles from the specified start location to the target location. Machine learning, a sub-field of artificial intelligence, is applied to create a Long Short-Term Memory (LSTM) neural network that is implemented and executed to allow a mobile robot to find the trajectory between two points and navigate while avoiding a dynamic obstacle. The input of the network is the distance between the mobile robot and the obstacles thrown by the LiDAR sensor, the desired target location, and the mobile robot’s location with respect to the odometry reference frame. Using the model to learn the mapping between input and output in the sample data, the linear and angular velocity of the mobile robot are obtained. The mobile robot and its dynamic environment are simulated in Gazebo, which is an open-source 3D robotics simulator. Gazebo can be synchronized with ROS (Robot Operating System). The computational experiments show that the network model can plan a safe navigation path in a dynamic environment. The best test accuracy obtained was 99.24%, where the model can generalize other trajectories for which it was not specifically trained within a 15 cm radius of a trained destination position.

Advancements in Learning-Based Navigation Systems for Robotic Applications in MRO Hangar: Review

“…The concept of robot operating system (ROS) navigation planners provides a detailed framework that can integrate different types of algorithms and sensors to develop complex robotics applications [ 112 ]. Emerging solutions have utilised neural networks (NNs) to optimise traditional path planning and obstacle avoidance techniques in recent times [ 43 , 113 ]. A neural network is a layered framework of interconnected nodes that takes input based on a designed task and produces network prediction or classification as output.…”

“…A neural network is a layered framework of interconnected nodes that takes input based on a designed task and produces network prediction or classification as output. The authors [ 113 ] applied a long short-term memory (LSTM) neural network using robot pose and agent to obstacle distance as the input to solve end-to-end path planning while avoiding a dynamic obstacle. Other authors applied multilayer perceptron (MLP) network classic path planning algorithms to demonstrate improved performance [ 114 ].…”

“…Most obstacle detection, avoidance, and path planning solutions are unstable and must be robust to be applicable in real-world scenarios. These systems rely on various parameters, and finding suitable configurations is difficult and results in unpredictable performance [ 113 ]. This can be exacerbated when dealing with limited training data.…”

Path planning for multiple agents in an unknown environment using soft actor critic and curriculum learning

“…Long 19 presented a multi‐scenario multi‐stage framework to learn a policy, which can achieve a sensor‐level collision avoidance. Molina‐Leal 20 adopted a long short‐term memory network to allow the robot to reach the goal without colliding with dynamic obstacles. Zhelo 21 adopted the asynchronous advantage actor‐critic (A3C) algorithm for mapless navigation with the sum of the external reward and the curiosity reward.…”

Optimization of an Autonomous Mobile Robot Path Planning Based on Improved Genetic Algorithms

“…The ability to navigate through a given environment is considered to be one of the most desirable characteristics of autonomous robots [5], [8]- [11]. Path planning is a method for an autonomous robot to get from the beginning point to the goal while traversing an environment that includes both static and dynamic obstacles [12], [13]. It is possible to divide path planning into global and local planning, depending on the scope of the map [8], [14]- [20].…”