Robust and reliable dual the classification system with error compensation module

• Main Author: Ahmad Nasiruddin, Ab Razak
• Format: Thesis
• Language: English
• Published: 2014
• Subjects: TJ Mechanical engineering and machinery
• Online Access: http://umpir.ump.edu.my/id/eprint/13947/1/FKASA_AHMAD%20NASIRUDDIN%20AB%20RAZAK.pdf; http://umpir.ump.edu.my/id/eprint/13947/

The toll industry put great emphasis on preventing revenue leakage especially from fraud by toll tellers. An automatic vehicle classification system (AVC) has been developed in order to mitigate the problem based on the classes defined by Malaysian Highway Authority (MHA). The system uses treadle sensors to map tires and count the number of axles while the optical barrier is used to detect the presence of vehicle. However, since the short lifecycle because of wear and tear due to the nature of them being treaded by the vehicles, the treadle breaks quickly and needs frequent replacement. Disruptions during maintenance and replacement causing inconvenience and unreliability that results in a steady demand by worldwide toll collection industries for replacement or at least enhancement of the technology. Two approaches have been made to mitigate this problem. First, the materials in the treadle are improved and second, a system to compensate for small errors is developed. The sensor is changed from fragile contact switch to strain gage that has high fatigue limit. The new sensor housing is made out of stainless steel that has high fatigue limit of 260 MPa. Furthermore, to increase the treadle endurance, the sensors are contained in elastomeric material which possesses hardness of 70 (Shore A scale) that protects the sensors from external impact of vehicles yet still allows force to be transferred to the sensor. The new treadle sensor was tested and can last more than 5 million treading cycles as required by the industry. Moreover, a robust dual tire classification module is installed in the system that can maintain high accuracy even when some sensors have error. Test using simulated data shows that the algorithm can maintain 100% accuracy when two errors are present. The module was also tested on actual data that gives 99.83% accuracy. With these enhancements, the lifecycle of the device is elongated beyond 5 million cycles without compromising the accuracy. Thus, it prevents revenue leakage and avoids traffic disruptions that will contribute positively to the economy.