POMDP-based Planning Model of Driving Force during Shield Tunneling

Abstract

In shield tunneling, the automatic control systems are mostly based on empirical relationships and do not have a precise theoretical background. Consequently, the present automatic control systems control the shield according to a snake-like motion to move it back onto the planned alignment. To solve such problems, an approach that adopts partially observable Markov decision processes(POMDP) theory to plan the driving force with explicit representation of the uncertainty during excavation is presented. In POMDP, the huge resistances during excavation, shield attitudes, and the driving force of possible world are scattered as the state set, observe set and action set, respectively. Furthermore, the deviations of shield attitude from planned alignment and the stationary of the loads acting on the shield are considered in the calculation of the value function. The experimental results confirm that the driving force planning based on POMDP is reasonable, and have certain ability of compliance at the mutation conditions.