"Optimal Sensor Fusion and Position Control of a Low-price Self-driving Vehicle in Short-term Operation Conditions"
- Year of publication
- J. Choi and K. KONG
- International Journal of Control, Automation and Systems
Real-time estimation of the absolute position and orientation plays a critical role in the control and path planning of a self-driving vehicle. Global positioning systems (GPSs) are used to estimate the absolute position of a vehicle from satellite signals, but their reliability and accuracy are not sufficient for precise control. An inertial navigation system (INS) is often used to complement the GPS, but it is always subjected to a drift problem due to integration. Such a problem has been solved by utilizing a high performance GPS, an INS, a vision sensor, a radar, a laser scanner, or their combination. However, these sensor systems are very expensive compared to the price of common commercial vehicles, which hinders the popularization of self-driving vehicles. Moreover, the target users of self-driving vehicles are most likely the elderly, as well as the blind, and thus the price is an important factor in the system design. In this paper, a sensor fusion method for a low-price self-driving vehicle and real-time path planning algorithm without a laser scanner are introduced. The proposed low-price sensor system consists of a GPS, an inertial measurement unit (IMU) with a three-axes accelerometer, a three-axes gyroscope, and a three-axes magnetometer, an encoder at the real wheels, and a potentiometer that measures the steering angle. The overall price is less than 300 USD. For the complete estimation of the absolute position and orientation of the vehicle, the proposed method estimates the driving velocity using a kinematic Kalman filter, then the orientation angle and absolute position are calculated by the recursive least squares method. The obtained information is then used for the real-time control of a self-driving vehicle. The proposed method is verified through simulation studies and by experimental results.