Fuzzy Logic Controlled Brakes for Trains Erik Lee May 2, 2012 EE Intelligent Algorithms Dr. Neil F. Palumbo

FUZZY LOGIC BASED AUTOMATIC BRAKING SYSTEM IN TRAINS G. Sankar and S. Saravana Kumar Bannari Amman Institute of Technology Sathyamangalam, Tamil Nadu, India Proceedings of India International Conference on Power Electronics 2006

Automatic Brakes leads to Minimized Cost and Time

Membership Functions 15 meters/s = 33.5 miles/h 40 meters/s = 89.5 miles/h 500 meters = 0.31 miles or 5.46 football fields

Rule Surfaces

Results Using 1.3 m/s^2 Deceleration km/h

Speed vs Distance Brake Deceleration 1.3m/s^2

Braking system Train weight with engine Decce leration (g) m/s^2 Rails Train speed (mph) Stopping distance (yards) Time to stop (s) Steel & McInnes air197t 7cwt Wet Smith vacuum262t 7cwt dry Clark and Webb chain 241t 10cwt dry Barker's hydraulic210t 2cwt dry Westinghouse vacuum204t 3cwt wet Fay mechanical186t 3cwt wet Clark hydraulic198t 3cwt dry Westinghouse automatic203t 4cwt dry Emergency Brake (Without track brakes) 1.5 Magnetic Track Brake1.8

Different Speeds 100 m/s = mph 60 m/s = mph (fastest)

Modified Rule Surface slope = slopeFunction(s); center = * s ^ 2; brakingSurface(sIndex, dIndex) = sigmoid(d, center, slope) * 100;

Modified Results

Modified Results Speed vs. Distance

Conclusion Fuzzy P.I.E. + Easier and Intuitive - Did not scale automatically with different speeds Modified controller + More accurate stopping from different speeds + Easier computation - Not intuitive and difficult to modify

Future Work Look into the grade of the track as an input