The uplink capacity of CDMA cellular networks is improved by adaptive sectoring based on tracking of mobiles' spatial distribution. The distribution is modeled as a spatial Poisson process, with its rate function quantizes the density of the active mobiles. The rate function's time dynamics is assumed to evolve according to mobiles' mobility pattern, and is formulated using the Influence model. In this paper, particle filtering is applied in the tracking and estimation of the mobile concentration based on network traffic, and it enables the computation of the network interference and thus the system outage probability. Different sectoring schemes are compared in terms of outage probabilities, and the minimum scheme is chosen for each time period. More specifically, the adaptive sectoring problem is formulated as a shortest path problem, and the optimal path corresponds to the sectoring scheme with the minimum outage probability.