Communication rates for tracking a moving target
Increasing data rates provides diminishing returns for tracking accuracy after a certain point.
Communication is not always beneficial for tracking. Very low rates can degrade performance because of coarse quantization.
Moderate rates recover most of the achievable gain. Increasing the rate from 4 to 12 bit/s per agent reduces the root-mean-square error (RMSE) from 13.93 m to 6.35 m.
A further increase to 32 bit/s per agent improves the RMSE only marginally to 5.93 m. This shows a knee region at moderate rates.
Stationary agents observe a moving target using heterogeneous range and bearing measurements. They perform local extended Kalman filtering (EKF).
Agents quantize their posterior state means. They fuse exchanged information via covariance intersection (CI).
The remaining loss after the knee region is mainly due to the conservatism of the fusion architecture. The benefit depends on the rate and sensing geometry.
Key facts
- Increasing the rate from 4 to 12 bit/s per agent reduces the RMSE from 13.93 m to 6.35 m.
- A rate of 32 bit/s per agent only improves the RMSE to 5.93 m.
- Very low communication rates can degrade performance due to coarse quantization.