Design and implementation of intelligent braking system

Abstract

Autonomous vehicles will be seen in the very near future on our roadways. When integrated with other advanced technologies like Global Positioning System (GPS), data communication system and so forth, a system with certain level of intelligence can be developed. The vehicles can be maneuvered automatically with highly optimized journey in terms of money, time, quality and efficiency. Intelligent braking system is one of vehicle subsystems that can be unprompted to function when needed. The system must be able to sense certain parameters like the speed of target vehicle and its own speed. From these measured variables the output brake force can be generated. To do this operation the system ought to have a knowledge base so as to have decision-making ability. The problem of modeling brake system is difficult due to its nature that inherits high nonlinear dynamics. It is possible, although laborious to come with analytical model of the system after severe constraints being imposed on the modeling. Technique of soft computing allows natural languages to play a role in scientific theory. It marks a significant paradigm shift, especially on how people see their analytical problems. Fuzzy logic can be an excellent tool to control a complex, nonlinear system. In this research, this tool is used to govern the brake system. Another aspect that has been tackled by this research work is the use of digital signal processor to run the fuzzy algorithm. It has several advantages over universal microcontroller and a dedicated fuzzy chip. The on chip peripheral allows the system to be composed in compact circuitry suitable for real time embedded system. It also provides room for other signal processing algorithms to be run within the same processor without any burning of cash needed to buy other processors. The result shows that the TMS320LF2407 processor can be a good processing engine for the braking system. It accords the simulation results with good accuracy. The time of Fuzzy Logic Inference Per Second (FLIPS) is also less than the time of scanning the input data.