Human drivers are remarkably robust against various unexpected occurring variations and corruptions by understanding temporal changes and traffic scenes. In contrast, the neural network based autonomous navigation system can be easily affected by sensor data anomaly, like occlusion, sensor noise, challenging weather and illumination conditions. Such external disturbances are inevitable in practical driving applications. In this paper, we develop a semi-supervised anomaly detection module to detect the corrupted data while extracting the traffic scenario features. We further introduce an end-to-end robust autonomous navigation framework based on the idea that the consecutive frames of clean data depict a similar traffic scenario and the differences among the sequential data imply the dynamic state changes. By taking into consideration both spatial traffic scenario and temporal environmental variation, the model is able to achieve robust navigation against sensor data corruptions. We conduct experiments in CARLA platform and the evaluation results show the effectiveness of the proposed method.

Recently, the Joint Video Exploration Team (JVET) has established the latest video coding standard, Versatile Video Coding (H.266/VVC). The H.266/VVC adopts the quadtree with nested multi-type tree coding block structure which brings huge computational burden that limits the development, adoption and application of the future video coding standard. To encounter this problem, we propose a novel fast block partition algorithm through Bayesian decision rule. The proposed method fully explored the information of the horizontal binary splitting which is newly introduced because of the adoption of the novel coding block structure. Experimental results demonstrate that the proposed algorithm achieves about 45% encoding time saving on average with negligible quality loss under all-intra configuration.

Recently, the Joint Video Exploration Team (JVET) has established the latest video coding standard, Versatile Video Coding (VVC). The VVC adopts the Multiple Transform Selection (MTS) as a supplement to the primary transform, where DST7 and DCT8 are included as the optional transform kernels. Though MTS provides appreciable coding gain, its computational complexity is relatively high since several transforms need to be evaluated through the Rate Distortion Optimization (RDO) process. To address this issue, we propose a novel fast two-stage algorithm that utilizes the spacial coding statistics and the primary transform information to early terminate the RDO process of MTS. Experimental results show that by employing this fast MTS method inside VVC, one can achieve up to 23% time reduction with only 0.16% Bjontegaard Delta bitrate (BD-rate) increase.