A PERCEPTION BASED ADVANCED VIRTUAL ROBOTIC ASSISTANCE TO DRIVER FOR TRAFFIC SIGN DETECTION

Now-a-days, due to the tremendous growth of vehicles, driving is become a complex and multitasking process which includes perception and cognition of drivers and motor movements. For safety concerns, the traffic signs and the vehicle information has to be displayed which strongly impacts the attention of drivers with reduced mental workloads. Drivers need some assistance to maintain their awareness and leading their attention to potential hazards. In this paper, a computer vision based technique is proposed to detect and recognize traffic signs based on their color and shape features. The investigation of traffic sign recognition has been of extraordinary attractions and the proposed system is addressed by a four phase system including segmentation, detection, feature extraction, and classification. We cover recent attempts to establish an automatic agent detect the traffic signs for vision based Driver Assistance System (DAS). We present a novel framework based on the system for recognizing circular, square, and triangular traffic signs makes utilization of artificial intelligence system, for example, heuristics functions to discover shapes. In order to prove the performance of the proposed system, the features of the detected traffic sign is given to three different classifiers such as j48, Naive Bayes, and k-nearest neighbor (k-NN) for classification and the obtained results are tabulated. The proposed framework is established in real world scenario and tested in highways