Impact of Perception Errors in Vision-Based Detection and Tracking Pipelines on Pedestrian Trajectory Prediction in Autonomous Driving Systems.

Pedestrian trajectory prediction is crucial for developing collision avoidance algorithms in autonomous driving systems, aiming to predict the future movement of the detected pedestrians based on their past trajectories. The traditional methods for pedestrian trajectory prediction involve a sequence of tasks, including detection and tracking to gather the historical movement of the observed pedestrians. Consequently, the accuracy of trajectory prediction heavily relies on the accuracy of the detection and tracking models, making it susceptible to their performance. The prior research in trajectory prediction has mainly assessed the model performance using public datasets, which often overlook the errors originating from detection and tracking models. This oversight fails to capture the real-world scenario of inevitable detection and tracking inaccuracies. In this study, we investigate the cumulative effect of errors within integrated detection, tracking, and trajectory prediction pipelines. Through empirical analysis, we examine the errors introduced at each stage of the pipeline and assess their collective impact on the trajectory prediction accuracy. We evaluate these models across various custom datasets collected in Taiwan to provide a comprehensive assessment. Our analysis of the results derived from these integrated pipelines illuminates the significant influence of detection and tracking errors on downstream tasks, such as trajectory prediction and distance estimation.

Supervised Object-Specific Distance Estimation from Monocular Images for Autonomous Driving.

Accurate distance estimation is a requirement for advanced driver assistance systems (ADAS) to provide drivers with safety-related functions such as adaptive cruise control and collision avoidance. Radars and lidars can be used for providing distance information; however, they are either expensive or provide poor object information compared to image sensors. In this study, we propose a lightweight convolutional deep learning model that can extract object-specific distance information from monocular images. We explore a variety of training and five structural settings of the model and conduct various tests on the KITTI dataset for evaluating seven different road agents, namely, person, bicycle, car, motorcycle, bus, train, and truck. Additionally, in all experiments, a comparison with the Monodepth2 model is carried out. Experimental results show that the proposed model outperforms Monodepth2 by 15% in terms of the average weighted mean absolute error (MAE).

Memristor-based spiking neural network with online reinforcement learning.

Neural networks implemented in memristor-based hardware can provide fast and efficient in-memory computation, but traditional learning methods such as error back-propagation are hardly feasible in it. Spiking neural networks (SNNs) are highly promising in this regard, as their weights can be changed locally in a self-organized manner without the demand for high-precision changes calculated with the use of information almost from the entire network. This problem is rather relevant for solving control tasks with neural-network reinforcement learning methods, as those are highly sensitive to any source of stochasticity in a model initialization, training, or decision-making procedure. This paper presents an online reinforcement learning algorithm in which the change of connection weights is carried out after processing each environment state during interaction-with-environment data generation. Another novel feature of the algorithm is that it is applied to SNNs with memristor-based STDP-like learning rules. The plasticity functions are obtained from real memristors based on poly-p-xylylene and CoFeB-LiNbO3 nanocomposite, which were experimentally assembled and analyzed. The SNN is comprised of leaky integrate-and-fire neurons. Environmental states are encoded by the timings of input spikes, and the control action is decoded by the first spike. The proposed learning algorithm solves the Cart-Pole benchmark task successfully. This result could be the first step towards implementing a real-time agent learning procedure in a continuous-time environment that can be run on neuromorphic systems with memristive synapses.