Multi-sensor data fusion using the influence model

System robustness against individual sensor failures is an important concern in multi-sensor networks. Unfortunately, the complexity of using the remaining sensors to interpolate missing sensor data grows exponentially due to the "curse of dimensionality". In this paper, we demonstrate that the influence model, our novel formulation for combining evidence from multiple interactive dynamic processes, can efficiently interpolate missing data and can achieve greater accuracy by modeling the structure of multi-sensor interaction