Position Control of a Mobile Robot through Deep Reinforcement Learning

Human-Guided Reinforcement Learning With Sim-to-Real Transfer for Autonomous Navigation

Reinforcement learning (RL) is a promising approach in unmanned ground vehicles (UGVs) applications, but limited computing resource makes it challenging to deploy a well-behaved RL strategy with sophisticated neural networks. Meanwhile, the training of RL on navigation tasks is difficult, which requires a carefully-designed reward function and a large number of interactions, yet RL navigation can still fail due to many corner cases. This shows the limited intelligence of current RL methods, thereby prompting us to rethink combining RL with human intelligence. In this paper, a human-guided RL framework is proposed to improve RL performance both during learning in the simulator and deployment in the real world. The framework allows humans to intervene in RL's control progress and provide demonstrations as needed, thereby improving RL's capabilities. An innovative human-guided RL algorithm is proposed that utilizes a series of mechanisms to improve the effectiveness of human guidance, including human-guided learning objective, prioritized human experience replay, and human intervention-based reward shaping. Our RL method is trained in simulation and then transferred to the real world, and we develop a denoised representation for domain adaptation to mitigate the simulation-to-real gap. Our method is validated through simulations and real-world experiments to navigate UGVs in diverse and dynamic environments based only on tiny neural networks and image inputs. Our method performs better in goal-reaching and safety than existing learning- and model-based navigation approaches and is robust to changes in input features and ego kinetics. Furthermore, our method allows small-scale human demonstrations to be used to improve the trained RL agent and learn expected behaviors online.

Deep Deterministic Policy Gradient-Based Autonomous Driving for Mobile Robots in Sparse Reward Environments

In this paper, we propose a deep deterministic policy gradient (DDPG)-based path-planning method for mobile robots by applying the hindsight experience replay (HER) technique to overcome the performance degradation resulting from sparse reward problems occurring in autonomous driving mobile robots. The mobile robot in our analysis was a robot operating system-based TurtleBot3, and the experimental environment was a virtual simulation based on Gazebo. A fully connected neural network was used as the DDPG network based on the actor-critic architecture. Noise was added to the actor network. The robot recognized an unknown environment by measuring distances using a laser sensor and determined the optimized policy to reach its destination. The HER technique improved the learning performance by generating three new episodes with normal experience from a failed episode. The proposed method demonstrated that the HER technique could help mitigate the sparse reward problem; this was further corroborated by the successful autonomous driving results obtained after applying the proposed method to two reward systems, as well as actual experimental results.

Detection of Static and Mobile Targets by an Autonomous Agent with Deep Q-Learning Abilities

This paper addresses the problem of detecting multiple static and mobile targets by an autonomous mobile agent acting under uncertainty. It is assumed that the agent is able to detect targets at different distances and that the detection includes errors of the first and second types. The goal of the agent is to plan and follow a trajectory that results in the detection of the targets in a minimal time. The suggested solution implements the approach of deep Q-learning applied to maximize the cumulative information gain regarding the targets' locations and minimize the trajectory length on the map with a predefined detection probability. The Q-learning process is based on a neural network that receives the agent location and current probability map and results in the preferred move of the agent. The presented procedure is compared with the previously developed techniques of sequential decision making, and it is demonstrated that the suggested novel algorithm strongly outperforms the existing methods.