-We are working to change the constants in the below variables in a hope to localize quicker and provide more accurate variable representation than what was originally given. -Steps include determining if the changed constant(s) still allows the robot to localize, and if so, how long it takes to localize. -This will easily allow us to determine which variables, when changed, perform better than what the current constants are already set at. -These constants were chosen for the tuning because there seems to be no explanation or metric for the values that are currently set, and so we decided that perhaps modifying them could enhance the robot's performance. *** -> Denotes changed variable constant ============ BASE TEST ============= # additional constants starting on line 230 of pf.py self.alpha1 = 0.2 self.alpha2 = 0.2 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Localizes relatively quickly and efficiently, but based on tests may sometimes be in the total wrong spot. AVG TIME TO LOCALIZE(when it did): 35 seconds ============= ============= TEST #1 ============= # additional constants starting on line 230 of pf.py *** self.alpha1 = 0.7 self.alpha2 = 0.2 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? No Has trouble finding where it is, does not have a definite spot where it thinks it is. TIME TO LOCALIZE: N/A ============= TEST #2 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.15 self.alpha2 = 0.2 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? No We drove it around for more than 2 minutes, it was extremely confident it knew where it was, but it was in the wrong spot the whole time NOTE: Map shape most likely has something to do with this TIME TO LOCALIZE: N/A ============= TEST #3 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.4 self.alpha2 = 0.2 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Still a decent spread of possibilities, but there is a good cluster at the correct location. This seemed to have an effect on localization. However, one of the iterations failed to localize at all. Iterations of Test: 3 AVG TIME TO LOCALIZE: 80 seconds ============= TEST #4 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.5 self.alpha2 = 0.2 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? No Went well over our established time threshold. TIME TO LOCALIZE: N/A seconds ============= TEST #5 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.3 self.alpha2 = 0.2 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? No Went well over our established time threshold. TIME TO LOCALIZE: N/A seconds ============= TEST #7 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.15 self.alpha2 = 0.2 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes A good number of outlying particles, but what seemed to be a pretty definitive idea of its location. TIME TO LOCALIZE: 34 seconds ============= TEST #8 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.25 self.alpha2 = 0.2 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Iterations of Test: 2 AVG TIME TO LOCALIZE: 35 seconds ============================================================================== ============= TEST #9 ============= # additional constants starting on line 230 of pf.py self.alpha1 = 0.2 ***self.alpha2 = 0.1 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes TIME TO LOCALIZE: 50 seconds ============= TEST #10 ============= # additional constants starting on line 230 of pf.py self.alpha1 = 0.2 ***self.alpha2 = 0.15 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Once localized, not many outlying particles and followed the robots real location almost perfectly. TIME TO LOCALIZE: 25 seconds ============= TEST #11 ============= # additional constants starting on line 230 of pf.py self.alpha1 = 0.2 ***self.alpha2 = 0.3 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Iterations of Test: 3 Once localized, not many outlying particles and followed the robots real location almost perfectly. TIME TO LOCALIZE: 18 seconds ============= TEST #12 ============= # additional constants starting on line 230 of pf.py self.alpha1 = 0.2 ***self.alpha2 = 0.4 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Iterations of Test: 3 AVG TIME TO LOCALIZE: 20 seconds ============= TEST #13 ============= # additional constants starting on line 230 of pf.py self.alpha1 = 0.2 ***self.alpha2 = 0.25 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Once localized, not many outlying particles and followed the robots real location almost perfectly. Iterations of Test: 3 AVG TIME TO LOCALIZE: 25 seconds ============= TEST #15 ============= # additional constants starting on line 230 of pf.py self.alpha1 = 0.2 ***self.alpha2 = 0.3 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Quickly localized. Iterations of Test: 3 AVG TIME TO LOCALIZE: 25 seconds =============================================================================== ============= TEST #16 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.15 ***self.alpha2 = 0.25 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Localized for the most part, but with a wide range of outlying particles and location possibilities. TIME TO LOCALIZE: 75 seconds ============= TEST #17 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.25 ***self.alpha2 = 0.15 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? No Was clustered, but in the wrong location (not even close) TIME TO LOCALIZE: N/A seconds ============= TEST #18 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.15 ***self.alpha2 = 0.15 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Localized quickly after some contest between two possible locations (we suspect helped by the route, which took the robot to the most angular portion of the map.) TIME TO LOCALIZE: 27 seconds ============= TEST #19 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.25 ***self.alpha2 = 0.3 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Identified quickly where it was on the map, thick particle cloud with few outliers in the robot's correct location. Iterations of Test AVG TIME TO LOCALIZE: 20 seconds ============= TEST #20 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.25 ***self.alpha2 = 0.25 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Identified quickly where it was on the map, pretty flawless with few outlying particles. Iterations of Test: 3 AVG TIME TO LOCALIZE: 15 seconds ============= TEST #21 ============= # additional constants starting on line 230 of pf.py ***self.alpha1 = 0.25 ***self.alpha2 = 0.4 self.alpha3 = 0.2 self.alpha4 = 0.2 self.z_hit = 0.5 self.z_rand = 0.5 self.sigma_hit = 0.1 CONCLUSION: LOCALIZES? Yes Took a little bit to get it to localize, but once it did it worked almost perfectly, with a thick particle cloud where the robot was and few outliers. Iterations of Test: 3 AVG TIME TO LOCALIZE: 40 seconds We find through this testing that values of .25/.25 and .25/.3 seem to have the best quality in the localization speed.