| China has the largest number of the blind in the world, how to walk safely is the biggest problem in the daily life of the blind. Road detection, obstacle recognition and target tracking have been the key technology in the field of guiding the blind, and are also the premises during obstacle avoidance process, so the mobile detection of obstacles has the important research value and engineering significance. Traditional tracking methods are basically based on the combination of multiple sensors, but have some problems like the disagreement of various experimental results. And the methods can only be applied in some specific situations. This paper proposes a new detection method with a single ultrasonic sensor and the improved extended Kalman filter. The paper focuses on the following aspects:1. This paper analyzes the relationship between the detection range and acoustic frequency, beam angle with the transducer radiation area which is a basis for the sensor selection. Three kinds of localization and tracking algorithms are introduced in detail. Two kinds of detection methods are put forward. The former combines EKF algorithm, the simulation analyzes the influence of the method; Basing on the other one, the EKF algorithm is improved. The simulation verifies the feasibility of the detection method and the performance of the improved algorithm.2. The mobile obstacle detection system is designed, including hardware and software. The hardware part introduces the work principle and the use method of each module. The software part takes the main program as the frame, and explains the working process of the four subroutines.3. The experimental platform is built, the measurement precision and the detection range of KS109 are studied in detail, and the stationary obstacles and moving obstacles are detected. The experimental results verify that the detection system has high measurement accuracy and stable performance, and can achieve the positioning and tracking of moving objects. |
PDF Full Text Request