Abstract
Path planning and obstacle avoidance are essential for autonomous driving cars. On the base of a self-constructed smart obstacle avoidance car, which used a LeTMC-520 depth camera and Jetson controller, this paper established a map of an unknown indoor environment based on depth information via SLAM technology. The Dijkstra algorithm is used as the global path planning algorithm and the dynamic window approach (DWA) as its local path planning algorithm, which are applied to the smart car, enabling it to successfully avoid obstacles from the planned initial position and reach the designated position. The tests on the smart car prove that the system can complete the functions of environment map establishment, path planning and navigation, and obstacle avoidance.
Highlights
In recent years, new industries such as cloud computing, big data, data center, virtual/augmented reality (VR/AR), 5G, artificial intelligence (AI), Internet of Things (IoT), and optical fiber sensing have emerged
The results show that the car can successfully avoid obstacles from the planned initial position and reach the designated position
SLAM technology is a very important part in the realization of intelligent obstacle avoidance vehicle technology. It is a technology for the intelligent vehicle to rely on various sensors to obtain information about the surrounding environment and establish a map of the environment, while using the created environment map to achieve reliable positioning
Summary
New industries such as cloud computing, big data, data center, virtual/augmented reality (VR/AR), 5G, artificial intelligence (AI), Internet of Things (IoT), and optical fiber sensing have emerged. The new intelligent obstacle avoidance vehicle uses a depth camera, which combines machine image recognition and machine vision technology It can fulfill lots of SLAM functions such as indoor mapping, automatic driving, and robotic arm grabbing on the robot operating system (ROS for short) [6]. Local path planning mainly includes artificial potential field method and dynamic window approach. The Dijkstra algorithm and A∗ algorithms are widely implemented in the ROS (robot operating system) [11] These methods are improved by reducing the number of searching grids through a heuristic estimation, the planning efficiency is inevitably low when the environment is complex, . Random Sampling Algorithms include BIT (batch informed trees), RABIT (regionally accelerated batch informed trees), RRT (rapidly exploring random tree), and Risk-DTRRT (risk-based dual-tree rapidly exploring random tree) These algorithms are more efficient and widely used in dynamic high-dimensional environments.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
