Adjacent-Lane Congestion Control for Stable Autonomous Driving

The patent segments the adjacent lane into multiple regions (first region closer to host vehicle, second region farther away) and applies different speed reduction strategies to each segment. When congestion is detected in the first region, the host vehicle speed is reduced by a first amount; when congestion is in the second region, speed is reduced by a second amount (different from the first). This segmented approach prevents collision while avoiding uniform deceleration that causes user discomfort.

The patent implements local quality by applying different speed control parameters to different spatial locations in the adjacent lane. The congestion state is determined separately for the first region and second region, and the host vehicle's driving speed is adjusted according to which specific region contains the congestion. This localized control strategy tailors the response to the actual spatial distribution of congestion, improving both safety and comfort.

The patent employs dynamic decision-making where the driving strategy is adjusted in real-time based on the detected congestion state. The controller dynamically selects between different speed reduction amounts (first amount vs. second amount) depending on which region contains the congestion. This dynamic adaptation achieves high driving accuracy without requiring a complex static system with pre-programmed responses for every possible condition.

The patent applies partial action by implementing a simplified two-region segmentation rather than dividing the entire lane into multiple fine-grained segments. This partial segmentation provides sufficient driving accuracy for safety purposes while keeping the system complexity manageable. The approach uses just enough complexity (two regions, two speed adjustment levels) to solve the problem without over-engineering.