Autonomous Vehicle Intent Projection System
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in determining and communicating their intended actions, especially in contexts involving behavioral communications at intersections, where human operators rely on gestures and cues to infer intentions.
Innovation Solution
The system tracks traffic objects approaching an intersection, determines their positions relative to the intersection, and generates a perceived right of way for the autonomous vehicle based on object types and behavior profiles, then presents an intent icon to the external environment proportional to the confidence in possessing the right of way, indicating whether to enter or remain stopped.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If autonomous vehicles use traditional signaling methods (turn signals, brake lights), then the communication capability is limited to basic signals, but the ability to convey complex intent and confidence levels is insufficient
Solution Approach 1:
The patent extends traditional 1D linear signaling (brake lights, turn signals) into 2D spatial communication by projecting icons at varying distances and positions. The distance and position of projected icons encode additional information dimensions, allowing the system to convey complex intent (direction, confidence level, timing) without adding more signal types, thus resolving the contradiction between information completeness and system complexity.
Solution Approach 2:
The system varies multiple parameters of the projected icons simultaneously: position (lateral and longitudinal), size, color, and distance from the vehicle. These parameter changes encode different aspects of intent communication, allowing a single projection system to convey rich information about the vehicle's planned maneuvers and confidence levels, addressing the limitation of traditional fixed-parameter signaling.
2Loss of information
If autonomous vehicles project intent icons at fixed distances, then the communication is simple to implement, but the ability to convey confidence levels and nuanced intent is limited
Solution Approach 1:
The system dynamically adjusts the projection distance and position of intent icons based on real-time confidence levels and contextual factors. Higher confidence results in projections at greater distances, while lower confidence results in closer projections. This dynamic adaptation allows the system to convey nuanced intent and confidence information without requiring complex additional hardware, resolving the contradiction between information richness and control complexity.
3Reliability
If autonomous vehicles do not communicate intent externally, then the system is simple, but human operators cannot infer the vehicle's intended actions, leading to safety issues
Solution Approach 1:
The patent introduces projected intent icons as an intermediary between the autonomous vehicle's decision-making system and human operators. These icons serve as a visual mediator that translates complex sensor data and decision algorithms into intuitive visual cues that humans can quickly interpret, improving safety by enabling effective human-vehicle communication without requiring complex direct interaction systems.
4Measurement precision
If autonomous vehicles process detailed traffic object behavior data, then the right of way determination is more accurate, but the computational complexity and processing time increase
Solution Approach 1:
The system extracts only the most relevant behavioral features from complex traffic object data (position, velocity, acceleration patterns, trajectory) rather than processing complete sensor datasets. This selective extraction maintains high accuracy in right of way determination by focusing on critical decision-making parameters while reducing computational complexity and processing requirements.
Data Source
AI summary
Systems and methods communicate an intent of an autonomous vehicle externally. In one implementation, traffic object(s) approaching an intersection towards which an autonomous vehicle autonomously navigates along a travel path is tracked. A position of the traffic object(s) relative to the intersection is determined when the autonomous vehicle is stopped at the intersection. A perceived possession of a right of way of the autonomous vehicle to enter the intersection is generated based on the position of the traffic object(s). A decision of whether to direct the autonomous vehicle into the intersection is generated based on the perceived possession of the right of way. An intent icon linked to an action of the autonomous vehicle is generated. The action is correlated with the decision of whether to direct the autonomous vehicle into the intersection. The intent icon is presented to an environment external to the autonomous vehicle.