This paper presents a lightweight stereo visionbased driving lane detection and classification system to achieve the ego-car’s lateral positioning and forward collision warning to aid Advanced Driver Assistance Systems (ADAS). For lane detection, we design a self-adaptive traffic lanes model in Hough Space with a maximum likelihood angle and dynamic pole detection ROIs, which is robust to road bumpiness, lane structure changing while the ego-car’s driving and interferential markings on the ground. What’s more, this model can be improved with geographic information system (GIS) or electronic map to achieve more accurate results. Besides, the 3D information acquired by stereo matching is used to generate an obstacle mask to reduce irrelevant objects’ interfere and detect forward collision distance. For lane classification, a Convolutional Neural Network (CNN) is trained by using manually labelled Region Of Interest (ROI) from KITTI dataset to classify the left/right-side line of host lane so that we can provide significant information for lane changing strategy making in ADAS. Quantitative experimental evaluation shows good true positive rate on lane detection and classification with a real-time (15Hz) working speed. Experimental results also demonstrate a certain level of system robustness on variation of the environment.

Deep reinforcement learning (DRL) has been successfully applied to end-to-end autonomous driving, especially in simulation environments. However, common DRL approaches used in complex autonomous driving scenarios sometimes are unstable or difficult to converge. This paper proposes two approaches to improve the stability of the training policy model with as few manual data as possible. For the first approach, reinforcement learning is combined with imitation learning to train a feature network with a small amount of manual data for parameters initialization. For the second approach, an auxiliary network is added to the reinforcement learning framework, which can leverage the real-time measurement information to deepen the understanding of environment, without any guide of demonstrators. To verify the effectiveness of these two approaches, simulations in image information-based and lidar information-based end-to-end autonomous driving systems are conducted, respectively. These approaches are not only tested in the virtual game world, but also applied in Gazebo, in which we build a 3D world based on the real vehicle model of Ranger XP900 platform, the real 3D obstacle model, and the real motion constraints with inertial characteristics, so as to ensure that the trained end-to-end autonomous driving model is more suitable for the real world. Experimental results show that the performance is increased by over 45% in the virtual game world, and can converge quickly and stably in Gazebo in which previous methods can hardly converge.

Autonomous driving, including intelligent decision-making and path planning, in dynamic environments (like highway) is significantly more difficult than the navigation in static scenarios because of the additional time dimension. Therefore, correlating the time dimension and the space dimension through prediction to create a spatio-temporal navigation map can make decision-making and path planning in such kinds of environment much easier. In this article, NGSIM data is analysed and processed from the perspective of the ego-vehicle (using the data as an ego-vehicle’s perception results). Based on the data, we develop an LSTM (Long-Short Term Memory)-based framework to predict possible trajectories of multiple surrounding vehicles within a certain range of the ego-vehicle. Then, the multiple predicted trajectories in a series of continuous dynamic highway scenes are projected into a spatio-temporal domain to create an octree map. Thus, dynamic targets and static obstacles can be unified into the same domain or map so that the dynamic disturbance problem for autonomous driving in highway environments can be resolved. Experimental results show that the proposed model is capable of predicting all the future trajectories around the ego-vehicle efficiently and the corresponding spatio-temporal map can be generated accurately in different dynamic scenarios.

The fusion of RGB and thermal images has profound implications for the semantic segmentation of challenging urban scenes, such as those with poor illumination. Nevertheless, existing RGB-T fusion networks pay less attention to modality differences; i.e., RGB and thermal images are commonly fused with fixed weights. In addition, spatial context details are lost during regular extraction operations, inevitably leading to imprecise object segmentation. To improve the segmentation accuracy, a novel network named spatial feature aggregation and fusion with modality adaptation (SFAF-MA) is proposed in this paper. The modality difference adaptive fusion (MDAF) module is introduced to adaptively fuse RGB and thermal images with corresponding weights generated from an attention mechanism. In addition, the spatial semantic fusion (SSF) module is designed to tap into more information by capturing multiscale perceptive fields with dilated convolutions of different rates, and aggregate shallower-level features with rich visual information and deeper-level features with strong semantics. Compared with existing methods on the public MFNet dataset and PST900 dataset, the proposed network significantly improves the segmentation effectiveness. The code is available at https://github.com/hexunjie/SFAF-MA.

The vision-aided inertial navigation is of vital importance to guarantee reliable navigation for unmanned ground vehicles (UGV) in GNSS-challenged environments. Yet, it is almost inevitable to occur unpredictable stochastic disturbances caused by ego-motion uncertainty due to uneven pavements, unstructured and dynamic scenes, which may lead to the divergence of estimation algorithms and the unreliability of positioning. In this paper, we aim at developing a robust strategy for vision-aided inertial navigation to resist uncertainties, and modifying unscented Kalman filter by formulating a novel adaptive factor according to the degree of abnormity (DoA). Applying the robust strategy and algorithms, a monocular camera and an inertial measurement unit are tightly fused to construct a robust vision-aided inertial navigation system. The robust method can automatically estimate the DoA online, and adaptively adjust the estimation error covariance matrix and the Kalman gain for different scenarios. As the modification is made in the framework of Kalman filter, the real-time performance, robustness and high-accuracy positioning can be achieved using the proposed adjustment strategy. Experiments with the KITTI dataset show that the robust vision-aided inertial navigation system possesses good robustness and self-adaptation for protection against ego-motion uncertainty, and can improve the positioning accuracy of unmanned ground vehicles.