eduzhai > Applied Sciences > Engineering >

Off-the-shelf sensor vs experimental radar -- How much resolution is necessary in automotive radar classification?

  • Save

... pages left unread,continue reading

Document pages: 8 pages

Abstract: Radar-based road user detection is an important topic in the context ofautonomous driving applications. The resolution of conventional automotiveradar sensors results in a sparse data representation which is tough to refineduring subsequent signal processing. On the other hand, a new sensor generationis waiting in the wings for its application in this challenging field. In thisarticle, two sensors of different radar generations are evaluated against eachother. The evaluation criterion is the performance on moving road user objectdetection and classification tasks. To this end, two data sets originating froman off-the-shelf radar and a high resolution next generation radar arecompared. Special attention is given on how the two data sets are assembled inorder to make them comparable. The utilized object detector consists of aclustering algorithm, a feature extraction module, and a recurrent neuralnetwork ensemble for classification. For the assessment, all components areevaluated both individually and, for the first time, as a whole. This allowsfor indicating where overall performance improvements have their origin in thepipeline. Furthermore, the generalization capabilities of both data sets areevaluated and important comparison metrics for automotive radar objectdetection are discussed. Results show clear benefits of the next generationradar. Interestingly, those benefits do not actually occur due to betterperformance at the classification stage, but rather because of the vastimprovements at the clustering stage.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...