A Bayesian approach to optimal sensor placement
International Journal of Robotics Research
Elements of information theory
Elements of information theory
Estimation of entropy and mutual information
Neural Computation
Near-optimal sensor placements in Gaussian processes
ICML '05 Proceedings of the 22nd international conference on Machine learning
A fuzzy neural network model for predicting clothing thermal comfort
Computers & Mathematics with Applications
| Hi-index | 0.00 |
Car cabins are transient, non-uniform thermal environments, both with respect to time and space. Identifying representative locations for the Heating, Ventilation and Air Conditioning (HVAC) system sensors is an open research problem. Common sensor positioning approaches are driven by considerations such as cost or aesthetics, which may impact on the performance/outputs of the HVAC system and thus occupants' comfort. Based on experimental data, this paper quantifies the spacialtemporal variations in the cabin's environment by using Mutual Information (MI) as a similarity measure. The overarching aim for the work is to find optimal (but practical) locations for sensors that: i) can produce accurate estimates of temperature at locations where sensors would be difficult to place, such as on an occupant's face or abdomen and ii) thus, support the development of occupant rather than cabin focused HVAC control algorithms. When applied to experimental data from stable and hot/cold soaking scenarios, the method proposed successfully identified practical sensor locations which estimate face and abdomen temperatures of an occupant with less than 0.7°C and 0.5°C error, respectively.