Key points are not available for this paper at this time.
Intelligent robots have the ability to self locate and build maps in unknown environments. This article designs a SLAM intelligent robot based on ROS, which collects environmental information, integrates lidar with depth camera, and implements motion control with STM32 as the core. Raspberry Pi is used as the upper computer, and the construction of 2D grid maps is completed using the SLAM algorithm Gmapping. The global and The local path planning adopts Adaptive Monte Carlo Localization (AMCL) to solve the overall route selection and dynamic obstacle problems of intelligent robots. In motion control, create motion control in ROS, achieve communication between Raspberry Pi and STM32, and subscribe to GeometryMsgs/Twist type speed related messages are used to select appropriate control strategies to drive the motor. Through RGB-D cameras, data acquisition and functional execution in path planning and obstacle avoidance are achieved, and relevant action instructions are transmitted to the main control microcontroller of the control section, thus achieving the execution of relevant actions. After comprehensive evaluation of system software and hardware testing, self positioning and map construction accuracy, as well as path planning and obstacle avoidance accuracy, the design of this system meets the requirements and can be widely applied in fields such as unmanned inspection, disaster rescue, and anti-terrorism and explosion prevention.
Lin et al. (Tue,) studied this question.