Design And Implementation Of Software-Hardware Cooperative Control System For Intelligent Arm

In recent years,China's e-commerce industry has developed rapidly,which has promoted the rapid development of express service industry.At present,one of the bottlenecks in the development of express service industry is the low level of intelligent infrastructure construction.With the increasingly close relationship between the ecommerce industry and express service industry,the bottleneck of the development of express service industry will inevitably hinder the further development of e-commerce industry.One of the measures to solve the problem mentioned above is to introduce intelligent equipment to improve the information level of express service industry.This paper designed a software-hardware cooperative control system for intelligent arm,which was applied to express sorting links.It makes the sorting and handling links of logistics industry more intelligent,and thus reduces the labor costs and enhances the competitiveness as well as influence of enterprises.The control system consists of vision sensor,controller with SOC chip and arm.The controller has two processing units,one is responsible for the task of "observation",that is,image acquisition,and the other is responsible for the task of "thinking",that is,target recognition and arm control.In the part of target recognition,background difference method was used to determine the range of the target,depth information extraction was used to solve the distance between the visual sensor and the target,SIFT(Scale-invasive feature transform)algorithm was used to extract the features of the target for matching,Meanshift clustering algorithm was used to cluster the matched feature points.In order to improve the performance of the system,the SIFT algorithm with high computational complexity was accelerated with FPGA,which constituted softwarehardware cooperation.In the feature descriptor extraction part of SIFT hardware acceleration module based on FPGA,by reducing the bit width of gradient information(including gradient magnitude and gradient direction),optimizing the generation of Gauss weighting coefficients,simplifying the calculation of trilinear interpolation coefficients,complex calculation such as exponent,trigonometric function and multiplication were avoided and reduced the complexity of hardware design as well as hardware resource consumption.Finally,the designed SIFT hardware module was verified and debugged on board of FPGA.Then it was added to the arm system for joint debugging.The experimental results showed that the software-hardware cooperative intelligent control system has better recognition effect and can greatly reduce the time of target recognition and has higher real-time performance compared with the pure software scheme.