Temporal Sensor Distortion Modeling for Object Attribute Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Temporal sensors often experience latency-related issues, leading to distorted representations of dynamic objects, which can affect the accuracy of object attributes and tracking in various applications.
Innovation Solution
The techniques described utilize distorted representations of objects in temporal sensor data to infer object attributes such as velocities, bounding boxes, and orientations, by comparing data from multiple sensors with different scanning directions or using machine-learned models to process time-dimensional sensor data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If temporal sensors are used to capture data over a period of time, then more complete information about dynamic scenes is obtained, but latency-related issues cause distorted representations of dynamic objects
Solution Approach 1:
The patent converts the harmful distortion caused by temporal sensor latency into a beneficial signal. By analyzing the distortion patterns in temporal sensor data, the system infers velocity information about dynamic objects. The machine-learned model processes the distorted representations to extract both the object attributes and the velocity caused by the latency-induced distortion, thereby transforming the harmful effect into useful velocity estimation.
2Measurement precision
If multiple sensors with different scanning directions are used, then velocity estimation accuracy is improved, but device complexity increases
Solution Approach 1:
The patent introduces a machine-learned model as an intermediary that processes sensor data from multiple sensors with different scanning directions. This intermediary learns to integrate the information from multiple sensors and extract velocity estimates, thereby managing the complexity of processing data from multiple sensors while improving velocity estimation accuracy.
3Measurement precision
If machine-learned models are used to process time-dimensional sensor data, then object attribute determination accuracy is improved, but computational requirements increase
Solution Approach 1:
The patent performs preliminary training of the machine-learned model offline using labeled temporal sensor data. This preliminary action prepares the model in advance so that during actual operation, the model can quickly process temporal sensor data and infer object attributes and velocities without requiring extensive real-time computation. The heavy computational lifting is done beforehand, reducing real-time energy consumption.
Data Source
AI summary
Techniques for determining attributes associated with objects represented in temporal sensor data. In some examples, the techniques may include receiving sensor data including a representation of an object in an environment. The sensor data may be generated by a temporal sensor of a vehicle and, in some instances, a trajectory of the object or the vehicle may contribute to a distortion in the representation of the object. For instance, a shape of the representation of the object may be distorted relative to an actual shape of the object. The techniques may also include determining an attribute (e.g., velocity, bounding box, etc.) associated with the object based at least in part on a difference between the representation of the object and another representation of the object (e.g., in other sensor data) or the actual shape of the object.