融合激光SLAM实现平衡车智能导航

Implementation of the intelligent navigation of balance vehicle with laser SLAM

国内现有的两轮智能平衡车几乎不具有自主避障与定位功能。为了提高其安全性与灵活性,在传统的平衡车基础上加入了激光SLAM技术,实现自主建图、路径规划、定位和避障。运用卡尔曼滤波器对六轴传感器采集的加速度和倾斜角进行数据融合处理,在建图和定位方面,采用Google发布的Cartographer算法,路径规划和避障上采用Navigation功能包集成的move_base下的Teb算法。由于激光雷达建图时需要车速比较缓慢,并且需要尽可能避免抖动,因此让车模保持稳定的运动状态就很重要,为此首先对传感器获得的数据进行滤波,其次对小车的PID参数进行细调。同时为了更方便地控制,加入蓝牙功能,通过蓝牙控制小车运动,实现快速建图。在加入了SLAM技术之后,传统的平衡车可以实现避障和定位功能,能够实时检测出静态和动态障碍物,并绕开障碍物规划出最优路线,实现了无人驾驶功能。

Based on the existing two-wheeled intelligent balance car in China,it has almost no autonomous obstacle avoidance and positioning functions.In order to improve its safety and flexibility,laser SLAM technology is applied to the traditional self-balancing car to realize autonomous mapping and path planning,positioning and obstacle avoidance.The Kalman filter is used to fuse the acceleration and tilt angle collected by the six-axis sensor.In terms of mapping and positioning,the Cartographer algorithm released by Google is used,and the Teb algorithm under move base integrated by the Navigation function package is used for path planning and obstacle avoidance.Since the speed of the vehicle is relatively slow and the jitter needs to be avoided as much as possible when mapping the lidar,it is very important to keep the vehicle model in a stable motion state.For this purpose,the data obtained by the sensor is firstly filtered,and then the PID parameters of the car are fine-tuned.At the same time,for more convenient control,the bluetooth function is added to control the movement of the car through bluetooth to achieve rapid map building.After adding SLAM technology,traditional self-balancing scooters can realize obstacle avoidance and positioning functions,detect static and dynamic obstacles in real time,and plan an optimal route around obstacles,realizing the function of unmanned driving.