fixed path planning. i was an idiot when i did the multiprocessing by assuming planner will run slower. it's faster. now we don't pile on the data_queue. this should be shared memory too and i don't know why i'm avoiding that like the plague
