Vehicle Tracking System Using Occultation Map Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing object tracking systems in motor vehicles fail to provide satisfactory tracking of moving objects, particularly during occultation phenomena where objects become hidden, leading to inaccurate predictions and potential safety hazards.
Innovation Solution
A method and system that involves observing targets, tracking their movements recursively, determining the type of target, and constructing an occultation map to predict movement and visibility, considering the type and width of targets to improve tracking accuracy and prevent misidentification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If optical sensors are used to monitor the vehicle's surroundings, then the system can detect visible obstacles, but the system cannot detect occulted objects when they are hidden by other objects
Solution Approach 1:
The system performs preliminary actions by predicting the future position and trajectory of occulted targets before they reappear. The recursive tracking module uses historical position data and motion models to forecast where hidden objects will be visible again, allowing the system to maintain continuous tracking awareness despite temporary occultation events.
Solution Approach 2:
The patent introduces an intermediary occultation map that mediates between the sensor observations and target tracking. This map represents the spatial occupancy of occluding objects and is used to infer the presence and position of occulted targets, allowing the system to reason about hidden objects through intermediate spatial reasoning rather than direct observation.
2Measurement precision
If the system tracks all moving objects without considering target type, then it can maintain simple tracking logic, but it cannot accurately predict movement during occultation or determine the correct number of targets
Solution Approach 1:
The system applies local quality by assigning different motion models and prediction parameters to different target types (pedestrian, vehicle, animal). Each target type has its own characteristic speed ranges, acceleration patterns, and occlusion behaviors. The target type determination module classifies objects based on observed motion characteristics, and this classification is then used to select appropriate prediction models for the recursive tracking module.
Solution Approach 2:
The patent utilizes parameter changes by dynamically adjusting tracking parameters based on target type classification. Different target types have different expected velocity ranges, turning radii, and occlusion probabilities. The system changes these parameters adaptively based on the classified target type, improving prediction accuracy without requiring a completely different tracking algorithm for each object category.
3Productivity
If the system allows acceleration when a tracked vehicle is hidden by another vehicle, then it can respond to apparent clear paths, but it compromises safety by not anticipating the hidden vehicle's reappearance
Solution Approach 1:
The system applies preliminary anti-action by using the occultation map to anticipate potential reappearance of occulted vehicles before they actually become visible. When the prediction module determines that a hidden vehicle is likely to reappear in the near future, it preemptively restricts acceleration commands, preventing the safety issue before it can manifest. This allows the system to maintain cautious behavior during occultation periods while still enabling acceleration when occultation is unlikely.
Data Source
Figure 1~2
Figure 3~5
Figure 6~7
AI summary
The invention relates to a method and a corresponding system for tracking moving objects. Said method for tracking target moving objects includes a first step of observing targets, a second recursive step of tracking the movements of said targets, and a third step of determining whether a target is concealed. The method also includes a recursive step of determining the type of target, the second step and the third step being implemented depending on the type of target.