Long Range Obstacle Distance Measurement Methods And Recognition Mechanism Research Via Ultrasonic Signal Of Tower Crane Anti-collision System

Tower cranes(hereinafter referred to as tower crane) is a symbol of crane lifting equipment with aerial operation, high center of gravity and wide working range in construction industry, it is widely used in horizontal and vertical lifting on construction site. With the continuous expansion of the scale of urban infrastructure, collisions probability of inter tower cranes and tower crane with surrounding obstacle increase, major accidents have occurred frequently and showing a rising trend. Regarding to the weakness of current tower crane anti-collision monitoring system and based on the thought of active obstacle detection and warning, research of long-range distance measurement using ultrasonic sensors for tower crane surrounding obstacle and feature of obstacle ultrasonic echo signals recognition methods were performed, it has important theoretical and engineering significance. The main research and innovative achievements of this paper can be summarized as follows:For the developed long-range ultrasonic distance measurement hardware devices of the project team, ultrasonic measurement system based on virtual instrument was constructed and the system optimization experimental study was carried out. PC software based on Lab VIEW was designed with acquisition and analysis capabilities of the ultrasonic signal. According to ultrasonic and vibration theory, the process of ultrasonic echo and the mechanism were analyzed. According to the experiments of number of transmitting ultrasonic excitation pulse to ultrasonic echo amplitude relations, the transmitting parameters of long-distance ultrasonic measurement system were optimized. Experiments of obstacle distance and ultrasonic incidence angle proved the obstacle detection validity of the measurement system within 30 meters range.As the traditional digital filter has a time delay of the peak coordinate and low signal to noise ratio which affects accuracy of distance measurement and target signal detection etc, improved wavelet threshold de-noising algorithm was proposed for analysis of non-stationary ultrasonic signals. De-noising evaluation was determined based on the characteristics of ultrasonic echo signal noise. Through experiments, the weakness of time delay, signal to noise ratio, mean square error and other aspects before and after the ultrasonic echo signal de-noising of traditional digital filtering algorithm were analyzed. According to the wavelet threshold de-noising methods, a new threshold function with parameter adjustable was proposed. The simulation and experimental research were performed with improved wavelet threshold de-noising algorithm, and the results showed that the improved threshold function has better de-noising effect and flexible adjustabililty; it can improve anti-interference capability of the detection system.Long-range distance measurement using the model of ultrasonic echo signal method was studied. Detailed simulation using the model and experimental device analysis of noise resistance, accuracy, stability and applicability of traditional threshold methods and cross correlation methods long-range distance measurement using ultrasonic sensors was performed. According to the weakness of these two methods, the relationship between waveform of ultrasonic echo force to rise phase and ultrasonic excitation pulse was studied; time-domain mathematical model of ultrasonic echo signal envelope curves was established, two improved methods that envelope peak method and improved peak combine envelope model method were proposed. Simulation and experimental results showed that long-range distance measurement performance of improved algorithms were better than traditional measurement methods, improved algorithms have been validated can meet the engineering precision requirements for both single and multi obstacles long-range distance measurement.Ultrasonic echo amplitude model and recognition mechanism of long-range distance obstacle were studied. Through theoretical and experimental analysis, the relationship between opening angles of ultrasonic and distance was performed,opening angle energy coefficient of different distances was obtained. The ratio between maximum effective circle area corresponding to different distance and the area of the obstacle was studied, the area scaling factor was obtained; the echo characteristics of angular, planar and column obstacles were analyzed. Considering all these factors, ultrasonic echo amplitude model of typical material and shape obstacle was established. The impact of incidence angle of ultrasonic sensor to time characteristics of echo vibration was studied; echo difference of different materials within the range of small dip of ultrasonic sensor was researched, and the echo features of time domain and time-frequency domain was extracted, material detection and classification was performed by using feature evaluation based on Euclidean distance criterion and support vector machine, the recognition results were super.