Bus Sensor Deployment for City Hot Spots Mobile Sensing

Vehicular mobile sensing is an emerging paradigm that employs sensor-equipped vehicles to collect data for environmental monitoring. Among various host vehicles, buses are particularly effective due to their high reliability, low operational costs, and extensive spatio-temporal coverage. Existing research on bus-based mobile sensing generally adopts an opportunistic sensing framework, wherein areas (or grids) are monitored according to assigned importance scores. However, this approach cannot guarantee regular and continuous monitoring of critical urban hot spots such as schools or stadiums. To address this gap, this paper proposes a sensor-bus allocation model designed to ensure consistent hot spot surveillance while simultaneously maximizing overall sensing quality. The model is formulated as a non-linear integer program where the objective is to enhance sensing quality, subject to constraints that enforce periodic hot spot coverage. We linearize the non-linear objective function using a piecewise affine approximation, enabling the model to be solved efficiently with commercial solvers. The effectiveness of the proposed model is validated through a real-world case study using bus network data from Chengdu, China.