In this work, we propose an approach for the autonomous navigation of mobile robots using fusion the of sensor data by a Double Deep Q-Network with collision avoidance by detecting moving people via computer vision techniques. We evaluate two data fusion methods for the proposed autonomous navigation approach: Interactive and Late Fusion strategy. Both are used to integrate mobile robot sensors through the following sensors: GPS, IMU, and an RGB-D camera. The proposed collision avoidance module is implemented along with the sensor fusion architecture in order to prevent the autonomous mobile robot from colliding with moving people. The simulation results indicate a significant impact on the success of completing the proposed mission by the mobile robot with the fusion of sensors, indicating a performance increase (success rate) of ≈27% in relation to navigation without sensor fusion. With the addition of moving people in the environment, deploying the people detection and collision avoidance security module has improved about the success rate by 14% when compared to that of the autonomous navigation approach without the security module.

This article presents a methodology for transmitting light detection and ranging (LiDAR) sensor data, which is frequently used in robots and autonomous vehicles, utilizing robot operation system (ROS) and LoRaWAN. A primary challenge is the limited bandwidth of the long range (LoRa) network, which restricts data transmission from high-volume sensors like LiDAR. To address this issue, we propose an edge processing technique that employs autoencoders to compress the LiDAR data into a latent representation for efficient transmission. Our method achieves an 82% reduction in the original LiDAR data size, allowing it to fit within constraints of LoRaWAN’s limited payload capacity. Additionally, our approach enables the reconstruction of the ROS topic on the receiver side, effectively extending the capability of ROS from local networks to wide-area networks (WANs). The results demonstrate the feasibility of LiDAR data transmission over LoRaWAN using our method, thereby supporting the deployment of LiDAR sensors in environments with constrained networks and limited mobile connectivity.

In this work, we propose an approach for the autonomous navigation of mobile robots using fusion of sensor data by a Double Deep Q-Network with collision avoidance by detecting moving people via computer vision techniques. We evaluate two data fusion methods for the proposed autonomous navigation approach: Interactive and Late fusion strategy. Both are used to integrate mobile robot sensors through the following sensors: GPS, IMU and, an RGB-D camera. The proposed collision avoidance module is implemented along with the sensor fusion architecture in order to prevent the autonomous mobile robot from colliding with moving people. The simulation results indicate a significant impact on the success of completing the proposed mission by the mobile robot with the fusion of sensors, indicating a performance increase (success rate) of 27% in relation to navigation without sensor fusion. With the addition of moving people in the environment, deploying the people detection and collision avoidance security module has improved about 14% the success rate when compared to that of the autonomous navigation approach without the security module. Video was developed with robot navigation using the DDQN-Late Fusion https://www.loom.com/share/684afa6a5b0148afadc9a200ab9f3483.

This article proposes an algorithm for autonomous navigation of mobile robots that merges Reinforcement Learning with Extended Kalman Filter (EKF) as a localization technique, namely, EKF-DQN, aiming to accelerate learning and improve the reward values obtained in the process of apprenticeship. More specifically, Deep Neural Networks (DQN - Deep-Q-Networks ) are used to control the trajectory of an autonomous vehicle in an indoor environment. Due to the ability of EKF to predict states, this algorithm is proposed to be used as a learning accelerator of the DQN network, predicting states ahead and inserting this information in the memory replay. Aiming at the safety of the navigation process, it is also proposed a visual safety system that avoids collisions of the mobile vehicle with people moving in the environment. The efficiency of the proposed algorithm is verified through computer simulations using the CoppeliaSIM simulator with code insertion in Python. The simulation results show that the EKF-DQN algorithm accelerates the maximization of rewards obtained and provides a higher success rate in fulfilling the proposed mobile robot mission compared to the DQN and Q-Learning algorithms.

Aplicações industriais de sistemas robóticos móveis autônomos tem exigido o desenvolvimento de algoritmos de navegação autônoma utilizando inteligência artificial. Este artigo propõe um método baseado em algoritmo de aprendizagem por reforço profundo e rede convolucional (DQN-CNN: Deep Q-Network with Convolutional Network) para propiciar navegação autônoma a robôs dotados de câmera RBG, onde a fase de treinamento deste algoritmo e realizada no ambiente de simulação Coppelia VREP com a placa de desenvolvimento Jetson Nano da NVIDIA. A metodologia proposta abrange desde o treinamento em ambiente de simulação até a integração com sistema real. Para prototipação, utiliza-se uma pista de navegação com o rob o controlado por algoritmo de aprendizado por reforço DQN. Os resultados apontam para efetividade da proposta, onde foi possível verificar o robo real desempenhando com sucesso a missão planejada. Espera-se que este trabalho possa contribuir para o avanço da área de robótica móvel e aprendizado por reforço, Além de fornecer uma metodologia útil para pesquisadores e desenvolvedores que trabalham com estas tecnologias.