Autonomous Trailer Path Planning Under Position Uncertainty
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles, especially larger ones like semi-trucks, face challenges in obstacle and collision avoidance systems due to the need for dedicated sensors on trailers, which is burdensome, costly, and reduces efficiency, especially when trailers lack compatible sensors, leading to position uncertainty issues.
Innovation Solution
The system generates and selects candidate paths for vehicles using an uncertainty representation, predicting the vehicle's location over time without requiring dedicated sensors on trailers, leveraging computations of propagated motion with disturbance considerations to model trailer position uncertainty during path planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If dedicated sensors are attached to the trailer for path planning and collision avoidance, then the reliability of obstacle detection is improved, but the device complexity and cost increase significantly
Solution Approach 1:
The patent creates a virtual model (copy) of the trailer's physical state and position uncertainty, using computational representations instead of physical sensors on the trailer. The system propagates motion models and uncertainty distributions to generate virtual trailer position data that feeds into path planning algorithms, eliminating the need for physical sensor copies on each trailer component.
Solution Approach 2:
The patent replaces the mechanical sensor system (physical sensors attached to trailers) with a computational/mathematical system. Instead of using mechanical or electronic sensors to detect trailer position, the system uses motion propagation models, uncertainty representations, and computational algorithms to estimate trailer position and generate safe paths, substituting physical detection with mathematical computation.
2Adaptability or versatility
If dedicated sensors are installed on every trailer for compatibility with tractor systems, then the adaptability of the autonomous vehicle system is improved, but the ease of manufacture and deployment deteriorates due to calibration requirements
Solution Approach 1:
The patent creates a universal computational framework that works with any trailer-tractor combination without requiring trailer-specific sensors or calibrations. The motion propagation model and uncertainty representation system are universally applicable to different trailer types, making the system multi-functional and adapter-less, where the same computational approach handles diverse trailer configurations.
Solution Approach 2:
The system enables trailers to effectively 'self-report' their position through computational propagation of motion models from the tractor's known state. Instead of requiring external sensors on trailers to measure and report position, the system uses the tractor's motion state and trailer kinematics to self-determine trailer position, eliminating the need for trailer-based sensing infrastructure.
3Measurement precision
If sensors are added to trailers that lack them, then the measurement precision of trailer position is improved, but the loss of time and efficiency increases due to installation requirements
Solution Approach 1:
The patent performs preliminary computational setup by establishing motion propagation models and uncertainty representations during system initialization. Instead of installing sensors on trailers before operation, the system pre-configures mathematical models that continuously compute trailer position based on tractor motion, performing the measurement setup action in advance through computation rather than physical installation.
Data Source
AI summary
In various examples, systems and methods are disclosed for generating and/or analyzing candidate paths for a multi-body vehicle—e.g., a tractor trailer truck—based on obstacle avoidance considerations and using an uncertainty representation for the vehicle. The uncertainty representation may correspond to a trailer portion of the multi-body vehicle to account for the variations in rotation of the trailer with respect to the tractor. As such, the uncertainty representation may be indicative of a probability that the trailer of the vehicle occupies locations and/or points in world space. This probability—combined with the probability of locations of actors in the environment—may be used to generate candidate paths that satisfy various constraints—e.g., a minimum stochastic distance—between the vehicle and the actor.