Zhao Dean

Tool Trajectory Optimization of Robotic Spray Painting

A model of paint deposition rate is established according to the experimental data. A complex curved surface has to be divided into several patches and trajectory optimization for each patch is performed. Optimization processes are developed to optimize the paint thickness for a surface with multiple patches.The pattern search method is adopted to optimize the parameters. A workpiece, which is a complex curved surface, is used to test the scheme. The results of experiments have shown that the trajectory optimization algorithm achieves satisfactory performance. This algorithm can also be extended to other applications.

Design of Traceability System for Pork Safety Production Based on RFID

With the development of society and the improvement of living standards, consumers’ demand for high quality meat products is constantly increasing.Traceability for livestock products is widely recognized to be an effective measure for any modern and integrated food safety control system. The quality tracing and traceability system of production’s entire processes is an important technology tool to protect pork safety. This paper proposes a RFID-enabled traceability system for pork supply chain.By adopting SQL Server 2000 databases and intelligent identification technology, a tracing system suitable to Chinese situation for monitoring and controlling quality of pork is constructed, and it manages to realize information traceability for entire pork production. The study indicates that the traceability system is valuable for practical reference and feasibly. It can help consumers to confide in pork safety and encourage the pork industry developing.

Remote Monitoring and Controlling System of Factory Aquatic Products breeding

The paper briefly introduces the constitution and the theory of operation of a computer-based remote monitoring and controlling system of factory aquatic products breeding with multi-environmental factors. This system is composed of many single-chip microcomputers used as fishponds controlling computers and two industrial control computers, one is used as remote monitoring computer and the other as field monitoring computer. It takes GPRS radio communication and Internet technology to communication between remote monitoring computer and field monitoring computer. The system can detect and control the breeding environmental factors such as temperature, content of dissolvent Oxygen, pH, water level, opacity and so on. It realizes the real-time monitoring and controlling the growth of the fish by computer image processing technology and it takes fault tree analysis technology to make the system reliability design. Once the system fault happens and the fish occurs pathological changes and unwell, the system can give an alarm in time and inform the manager automatically by short messages.

Research on PID controller based on RBF neural network

RBF neural network is a kind of feedforward neural network with good performance, which is proposed based on the brains neurons to external reaction and it has the capability of strong nonlinear mapping, it can with arbitrary precision approaches a nonlinear function global, it has the advantage of simple structure, fast learning etc. Neural network with arbitrary nonlinear expression ability can realize the best PID control by system performance study, by RBF network, the self-learning PID controller can been built with parameters Kp, Ki, Kd, the simulation results show that the system has good dynamic and static performance.

Prediction system of sewage outflow COD based on LS-SVM

To solve the multi-variable, nonlinear and large time delay problems in the sewage treatment process, a prediction model of sewage outflow COD based on the Least Square Support Vector Machine (LS-SVM) is proposed. By converting the inequality constraints into equality constraints, the model transforms solving the SVM from a Quadratic Programming (QP) problem to a group of linear equations, which simplifies the learning process and improves the calculating efficiency. Compared with BP neural network, the experimental results verify that LS-SVM method has effectively improved performance in predicting sewage outflow COD. Some researches on empirical application have been done with the monitoring data in a wastewater treatment plant to verify the effectiveness and feasibility of the model.

Research on PID controller based on the BP neural network

The BP neural network is a multilayer feedforward network which spreads error inversely, the BP network can learn and store a lot of input/output mapping relationship without prior reveal the mathematical equations. The learning rule is to use the steepest descent method, the weights and threshold of network are adjusted constantly by the back propagation, which makes the network error squares minimum. Neural network with arbitrary nonlinear expression ability can realize the PID control which has the best combination by studying system performance, by the BP network, the parameters Kp, Ki, Kd self-learning PID controller can been built, the simulation results show that the system has good dynamic and static performance.

Using MATLAB image processing to monitor the health of fish in aquiculture

A particular and real-time image processing application is proposed to detect pathological changes to fish. Images are periodically acquired from a video source using specialized control software. Then, analog signals are converted into digital signals by image collection card. According to the theory of shrink and expand, binary image is used to eliminate impurity. After that, use the concept of area-square to find out the changes of pathological fish, that is the area of the white pixel. Thus, as long as counting out the number of white pixel and comparing with the statistical data, the condition of the fish will be immediately found out. And the processed results can be returned to control software and displayed in real time.

The research on target recognition and image tracking system based on high-speed DSP

After image tracking having been applied in many fields, people are now paying more and more attention to the research in finding a real-time image tracking system with preferable precision. This system is designed to be a master-slave structure based on PC and TMS320C6201(DSP). With the DSP focusing on image processing a real-time tracking system can be realized. The classical pattern arithmetic ATR is applied in the tracking. In the paper an effective hybrid genetic algorithm (GASA) is adopted in image segmentation. The effect of image segmentation is achieved by combining the capacity of GA to reach the global optimum with the capability of SA to gain the local search ability. A correlation tracking mode optimized by SSDA can improve the tracking probability.

A nighttime image enhancement method based on Retinex and guided filter for object recognition of apple harvesting robot

In order to improve the working efficiency of robot promptly picking ripe apples, the harvesting robot must have the ability of continuous recognition and operation at night. Nighttime apple image has many dark spaces and shadows with low resolution, and therefore, a Retinex algorithm based on guided filter is presented to enhance nighttime image in this article. According to color feature of image, the illumination component is estimated by using guided filter which can be applied as an edge-preserving smoothing operator. And the reflection component with image itself characteristics is obtained by employing single-scale Retinex algorithm. After gamma correction, these two components of image are synthesized into a new enhanced nighttime apple image. Fifty nighttime images acquired under fluorescent lighting are selected to make experiment. Experimental results show that the image enhancement performance indexes processed by the proposed algorithm, such as average gray value, standard deviation, information entropy, average gradient, and segmentation error are superior to those of histogram equalization algorithms and Retinex algorithm based on bilateral filter. In addition, compared with the Retinex algorithm based on bilateral filter, the proposed algorithm has an average reduction of 74.56% in running time with better real-time and higher efficiency. So it can realize the continuous operation of apple harvesting robot at night.

Obstacle modeling and information description method based on C-space for harvesting robot

For meeting the real-time working requirement of harvesting robot, the obstacles in the picking environment were equivalent to point-shaped, round-shaped and line-shaped obstacles according to their growth characteristics in the nature environment, which can avoid a great deal of computation and decrease the process time. And then the corresponding obstacles modeling method in C-space was presented according to the inherent geometric characteristic of the mechanical arms and the collision conditions between the mechanical arms and obstacles. Lastly, the obstacle information description method in C-space was determined by discussion and comparison for facilitate the obstacle avoidance path planning.