Author :
Li, Xi-ping ; Gu, Li-chen ; Jia, Jia
Author_Institution :
School of Mechanical and Electrical Engineering, Xi´an University Of Architecture And Technology, 710055, China
Abstract :
Recently, tower crane safety focused on anti-collision problem as construction environment and group of tower crane got more and more complex. The mainstream anti-collision methods used at domestic and overseas was still passive security model represented by working area limitation technology, which lacked flexibility, initiative and portability. Other on-going researches which adopted advanced sensors, communication techniques, intelligence algorithm couldn´t use in construction practice successfully. This paper presented innovatively a practical methodology which used ultrasonic sensors to gradually obtain the accurate state of barriers, security judgment and Anti-Collision alarm through multi-sensor information fusion techniques such as Kalman optimal fusion estimation, neural networks time fusion, fuzzy clustering space fusion and curve fitting prediction, etc. It was verified successfully that the method could sense distance with a wide range, rapidity detect either multiple types of stationary or moving objects, accurately determine the surface outline and position of each stationary barrier as well as the track of each moving objects via simulations and experiments, so that can give operators sufficient time to take corrective action in alarm mode before a collision occurs. Moreover, this methodology had characteristics of low cost, high precision, real-time.