FSDrive introduces a spatio‑temporal chain‑of‑thought approach that enables visual language models to generate future driving scenes as images, improving trajectory planning accuracy and safety by eliminating cross‑modal gaps and enforcing physical constraints in autonomous driving.
Tencent’s new wheel‑legged robot Ollie combines a compact 35 cm frame with advanced nonlinear control, whole‑body dynamics and trajectory planning to perform high‑speed skating, 360° flips and impressive jumps, a breakthrough showcased at ICRA2021 and promising future service applications.
The behavior prediction module in autonomous driving forecasts surrounding agents’ future actions using kinematic, map‑based, and machine‑learning methods, models multimodal uncertainty, and informs planning to adopt safer, conservative maneuvers, while research seeks richer features, rare behavior handling, and improved uncertainty representations.
This article presents a comprehensive overview of autonomous‑vehicle trajectory planning, covering its fundamental concepts, optimization formulation, decision‑making strategies, lateral and longitudinal planning methods, and the practical challenges faced in real‑world deployments.
This article explores how autonomous driving, driven by artificial intelligence, can dramatically improve safety, convenience, efficiency, and reduce congestion, outlines the five SAE levels, describes the three-layer control architecture, and explains key AI tools such as occupancy grids and cones of uncertainty that enable precise trajectory planning.
