COMPRESSION ALGORITHMS FOR VEHICLE MEASUREMENT OF PREVIEW DATA

摘要

Disclosures teach providing data over a vehicle bus, the data providing values for a road attribute indexed to points that are expected to be traversed by a vehicle. To compress the data, a server can adapt a regression function to the data, using coefficients on kernels in the function. The server may limit the number of coefficients to reflect a limitation on the rate of change imposed on the data from the road being reported. Similarly, the server may limit one or more numbers of bits encoding the coefficients in a message transmitted over the vehicle bus, reflecting a limited range of possible amplitudes imposed on the road for data. A client can receive a copy of the kernel and retrieve the coefficients from the message. The client may estimate future values of the road attribute by evaluating the coefficients applied to the kernels, which in some examples may be Gaussian.