April 2018
Intermediate to advanced
334 pages
10h 18m
English
Recurrent temporal aggregation involves aggregating environmental states across different time steps. Let's discuss the reason behind this in detail. First, fetching environmental states is not an easy task and sensor readings provide the best possible state representation of the environment. Therefore, state information of the current time step is not enough to get the full information of the environment. Therefore, integration of state information over multiple time steps captures the motion behavior, which is very important in the case of autonomous driving where the environmental state changes in split seconds.
Thus, by adding recurrence, handling of POMDP (partially observable Markov decision process) ...