Peer-Assisted Safety Models for Sensor-Limited Light Vehicles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous and assisted-driving vehicles, particularly lightweight two-wheelers and three-wheelers, face challenges due to limited sensor arrays and computation resources, making existing safety models designed for heavier vehicles ineffective.
Innovation Solution
Lightweight vehicles leverage sensing and computation capabilities from more well-equipped heavyweight vehicles and infrastructure elements to receive and transform safety models tailored to their specific kinematic and sensor contexts, enhancing safety through peer-assisted models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing safety models designed for heavier vehicles are used for lightweight vehicles, then the safety models can provide comprehensive safety constraints, but they become ineffective due to limited sensor arrays and computation resources of lightweight vehicles
Solution Approach 1:
The patent transforms the universal safety model into vehicle-specific safety models by applying local quality adjustments. The system generates customized safety models tailored to each vehicle type's specific sensor array capabilities and computation resources, making the safety constraints locally optimized for each vehicle category rather than universally applied.
Solution Approach 2:
The patent changes key parameters of the safety model including sensor array configurations, computation resource allocations, and safety constraint thresholds to match the specific characteristics of different vehicle types. This parameter transformation enables the safety model to adapt from heavy vehicle specifications to lightweight vehicle specifications.
2Device complexity
If lightweight vehicles use their own limited sensor arrays and computation resources, then they maintain system simplicity, but they cannot execute comprehensive safety models effectively
Solution Approach 1:
The patent introduces an intermediary transformation process that adapts comprehensive safety models into simplified versions suitable for lightweight vehicles. This intermediary system translates complex safety constraints into forms that can be executed by vehicles with limited sensor arrays and computation resources, bridging the gap between model comprehensiveness and system capabilities.
Solution Approach 2:
The patent creates simplified copies of comprehensive safety models that are tailored to lightweight vehicle capabilities. Instead of attempting to run the full comprehensive model, the system generates adapted copies that capture essential safety constraints while being compatible with limited hardware resources.
3Reliability
If safety models are customized for each vehicle type, then safety effectiveness is improved, but the complexity of generating and managing multiple customized models increases
Solution Approach 1:
The patent creates a universal model generation system that can produce multiple vehicle-specific safety models from a single comprehensive template. This multi-functional system handles transformation for different vehicle types (motorcycles, scooters, rickshaws, etc.) using a unified approach, reducing the complexity of managing custom models for each vehicle type.
Solution Approach 2:
The patent performs preliminary transformation of comprehensive safety models into vehicle-specific versions before deployment. By pre-processing and adapting safety models in advance for different vehicle categories, the system reduces the computational burden during real-time operation and simplifies model management through advance preparation.
Data Source
AI summary
Disclosed herein are systems and methods for peer-assisted safety models for autonomous and assisted-driving vehicles. In an embodiment, a safety-model service receives a safety-model request for a safety model from a target vehicle. The safety-model service identifies, responsive to receiving the safety-model request, one or more source vehicles as safety-model input sources. The safety-model service receives safety-model data associated with the identified one or more source vehicles. The safety-model service generates, based on the safety-model request and the received safety-model data, a target-vehicle safety model for the target vehicle. The safety-model service transmits the target-vehicle safety model to the target vehicle for use by the target vehicle.