Implementation of Industrial Object Recognition Robot using Optimized HOG& Cost efficient IoT System

Abstract

To enhance the performance of an industrial robot by optimizing an object recognition mechanism of a robotic module using advanced modified HOG algorithm and real time data transmission in the IoT perspective with reduced delay time. The object recognition process of an robotic module starts with capturing the target object, comparing with the database object and finding the matched object by modified HOG algorithm using MATLAB. Monitoring & Object recognition has been made live using IoT network with low power consumption. The user should be able to monitor and control the robotic module online from any part of world through internet. Our main objective is to increase the Accuracy Rate to 97.8% by optimizing HoG algorithm using LBP & SVM for object recognition. Next, decrease Delay Jitter to 25 to 33ms by optimizing transmission Protocol, Bandwidth (upstream & downstream),and priority configuration to increase transmission speed via internet. At last, cost efficient implementation is done by real-time optimization considering power consumption, hardware and software used for cloud computing, security and storage cost, hence overall cost of implementation is reduced to 30-40%.