After 2015 Taiwan Micromouse and Intelligent Robot Contest, I did a new mouse. In fact, In the hardware section, my old mouse and new mouse are not too different. I just increase a vacuum fan and remove 4 ABS cover which lock the IR sensors because I want to reduce mouse weight and lower gravity. About the vacuum fan, when I try fast run, I always control PWM at 80% about 4 voltage, and it give my mouse about 75 gram suction. I think It is enough run and have a good effect.
In the Algorithm section, I rewrite a new algorithm after The 35 All Japan Micromouse contest. I change some rule and increase speed value, my mouse can run the better path than before. I fill the different flood value at maze edge, such as running at straight or running at diagonal. If the direction is the same, I give it a reward value(reduce flood value). If the direction different as last, I will give it a punishment value(increase flood value). This is my method, I think it is easy to do. However, I have a bug at this contest. I want to find the best path, I set the goal and change each at center and start. If mouse think it searches finish, it will go back. Sometimes, It will change the goal each and do it, do it and do it… However, I think this is a good acquisition, I can find some different bug and revise by joining contest. It is very interesting for me.