Remote Function Selection Using Action Prediction Confidence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated drive vehicles face challenges in selecting the most appropriate remote function to execute when multiple remote functions are available, particularly in situations where the action prediction confidence degree for targets around the vehicle is low or the relative distance between the vehicle and targets is significant, affecting the efficiency and safety of remote operation.
Innovation Solution
A remote function selection device is configured with an automated drive determination unit, a remote function determination unit, a confidence degree calculation unit, and a function selection unit that calculates an action prediction confidence degree and selects the remote function based on this confidence degree and relative distance, opting for longer operator involvement when confidence is low and shorter involvement when distance is greater.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the automated drive vehicle selects a remote function based on basic executability, then the remote operation can be initiated, but the selection appropriateness deteriorates when multiple functions are executable
Solution Approach 1:
The patent introduces new parameters (action prediction confidence degree and relative distance) to the selection process. When multiple remote functions are executable, the system calculates these parameters for each function and selects based on the combined assessment, transforming a simple executability check into a multi-parameter optimization problem that resolves the contradiction between selection quality and system complexity
Solution Approach 2:
The patent replaces manual or simple automated selection mechanisms with an intelligent decision-making system that uses confidence degree calculation and distance-based filtering. This substitution enables more sophisticated selection criteria without proportionally increasing operational complexity, as the system automatically processes these parameters through defined algorithms
2Reliability
If the vehicle uses simple remote function selection, then the operation speed is fast, but the safety deteriorates when action prediction confidence is low
Solution Approach 1:
The system performs preliminary calculations of action prediction confidence degrees and relative distances before the actual remote function selection. By pre-computing these parameters and establishing selection criteria in advance, the system minimizes real-time decision delays while ensuring safe selections, even when confidence levels are low
Solution Approach 2:
The automated drive vehicle independently evaluates multiple remote functions using its own sensors and algorithms to determine confidence degrees and distances. This self-service capability allows the vehicle to perform comprehensive safety assessments without external intervention, maintaining both safety and operational speed
3Productivity
If the vehicle selects remote function without considering action prediction confidence, then the selection process is simple, but the operational efficiency deteriorates
Solution Approach 1:
The patent applies different selection criteria to different remote functions based on their specific characteristics. Each function is evaluated with respect to its own action prediction confidence degree and relative distance metrics, allowing the system to optimize for efficiency in each local context rather than applying a uniform complex algorithm to all functions
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A remote function selection device (24) selects a remote function in an automated drive vehicle configured to execute automated drive and remote travel and provided with a plurality of remote functions. The remote function selection device (24) is configured to determine whether executing the automated drive at a predetermined timing is impossible; determine the remote function that is executable at the predetermined timing when it is determined that executing the automated drive is impossible; predict an action of a target around the automated drive vehicle based on a detection result from an external sensor configured to detect the target; calculate an action prediction confidence degree for the predicted action of the target; and select the remote function. The remote function selection device (24) is configured to select the remote function based on the calculated action prediction confidence degree when it is determined that a plurality of remote functions is executable.