A METHOD FOR SYNTHETIC SENSOR DATA GENERATION

The herein disclosed technology relates to synthetic sensor data generation. More specifically, the disclosed technology at least partly relates to a computer-implemented method (100) for synthetic sensor data generation using a scenario database, wherein the scenario database comprising a plurality of scenario samples, each scenario sample comprising sensor data depicting a surrounding environment of a vehicle over a period of time, wherein each scenario sample is associated with a scenario embedding representing the scenario sample in a multi-dimensional space, wherein each scenario embedding is associated with a transformation volume in the multi-dimensional space, wherein the transformation volume is indicative of a set of possible transformed scenarios that can be generated from the corresponding scenario sample. The method (100) comprising: obtaining (5102) a request specifying a query scenario, wherein the query scenario is associated with a query embedding representing the query scenario in the multi-dimensional space; identifying (S106), within the scenario database, at least one scenario sample having a transformation volume within which the query embedding is located; and in response to successfully identifying the at least one scenario sample, generating (5108) synthetic sensor data corresponding to the query scenario by transforming the sensor data of the at least one identified scenario sample to synthetic sensor data having an associated scenario embedding in the multi-dimensional space within a threshold distance from the query embedding.