Occluded Object Detection via Remote Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle sensors, such as LIDAR, cameras, and radar, often fail to detect objects that are occluded by obstacles like buildings or are beyond their sensing range, leading to incomplete environmental maps and potential collisions in autonomous driving systems.
Innovation Solution
A vision system within the vehicle acquires remote sensor data from other vehicles and devices to supplement local sensor data, identifying occluded objects and rendering their location on a display to improve situational awareness and path planning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the vehicle uses its own sensors to build a map of the surrounding environment, then the vehicle can navigate and avoid objects within sensing range, but the map does not provide a complete representation of the surrounding environment due to occluded objects
Solution Approach 1:
The patent introduces remote sensors from other vehicles and devices as intermediaries to provide information about occluded objects. These remote sensors act as mediators that allow the vision system to detect objects beyond the vehicle's direct sensing range without requiring the vehicle to have omnidirectional sensors itself.
Solution Approach 2:
The patent makes the vehicle's vision system multi-functional by enabling it to both process local sensor data from its own sensors and integrate remote sensor data from external sources. This universal approach allows a single system to achieve both functions: direct detection and indirect detection of occluded objects.
2Reliability
If the vehicle operates in autonomous mode using only its own sensors, then the vehicle can perform path planning and obstacle avoidance, but the vehicle may encounter unforeseen obstacles causing erratic maneuvers
Solution Approach 1:
The patent applies preliminary action by acquiring remote sensor data before the vehicle would otherwise encounter occluded objects. The vision system proactively gathers information about hidden objects from remote sensors, allowing the autonomous driving module to plan paths in advance while being aware of potential obstacles that are not yet in the direct sensing range.
Solution Approach 2:
The patent implements feedback by continuously updating the environmental model with remote sensor data about occluded objects. This feedback loop allows the autonomous driving module to adjust path planning and obstacle avoidance strategies based on information about hidden objects, preventing erratic maneuvers caused by unexpected encounters.
3Length of stationary object
If the vehicle sensors have extended sensing range, then the vehicle can detect objects farther away, but objects beyond a particular distance still cannot be perceived
Solution Approach 1:
The patent uses copying by acquiring sensor data from remote sensors that are physically located near occluded objects. Instead of relying solely on the vehicle's own sensors to detect distant objects directly, the system copies detection data from remote sensors that have line-of-sight to these objects, effectively extending the sensing range without physical sensor extension.
Data Source
AI summary
System, methods, and other embodiments described herein relate to acquiring remote sensor data and identifying occluded objects from the remote sensor data. In one embodiment, a method includes acquiring, within a vehicle from one or more remote sensors, remote sensor data that indicates information about a nearby environment of the vehicle. The method includes identifying an occluded object in the surrounding environment according to the remote sensor data, and local sensor data generated by the vehicle. The occluded object is obscured from being perceived by one or more vehicle sensors of the vehicle. The local sensor data indicates information about at least nearby objects that have been observed by the vehicle sensors. The method includes rendering, within the display of the vehicle, a visualization of the occluded object to provide a view of the occluded object when the occluded object is otherwise obscured from being perceived by the vehicle.