Sensor Translation in Heterogeneous Arrays for Failure Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous or semi-autonomous systems relying on heterogeneous sensor arrays face operational failures when one sensor malfunctions, leading to complete loss of capabilities due to dependency on specific sensor types, necessitating a safe stop or landing protocol.
Innovation Solution
A perception engine aggregates data from multiple sensors to detect deficiencies and uses a sensor translation model to convert data from functioning sensors into synthetic data for the malfunctioning sensor, maintaining operational capabilities without redundant sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the autonomous system uses a heterogeneous sensor array with specific sensor types for specific algorithms, then the measurement precision and detection capability are improved, but the reliability deteriorates when a sensor malfunctions
Solution Approach 1:
The patent creates synthetic sensor data as a copy of the missing sensor's output by translating data from available sensors. When a sensor fails, the system generates synthetic data that replicates what the failed sensor would have detected, maintaining the data flow to dependent algorithms without requiring physical redundant sensors
Solution Approach 2:
The patent introduces a sensor translation model as an intermediary component that converts data from available sensors into the format and characteristics of the failed sensor type. This intermediary layer allows algorithms to receive data in their expected format even when the original sensor is unavailable
2Object-affected harmful factors
If the system programs autonomous apparatuses to stop or land safely when a sensor fails, then the safety is improved, but the productivity deteriorates due to operational interruptions
Solution Approach 1:
The patent performs preliminary action by pre-training sensor translation models during system operation. When sensors are functioning normally, the system learns translation mappings between different sensor types, so that when a sensor fails, the pre-learned translation capability can immediately generate synthetic data without interrupting operations
Solution Approach 2:
The patent ensures continuous operation by maintaining data flow to all algorithms even when sensors fail. The sensor translation model continuously generates synthetic sensor data from available sensors, allowing the autonomous system to maintain its navigation and detection capabilities without stopping
3Reliability
If the system uses redundant sensors to compensate for sensor failures, then the reliability is improved, but the device complexity and cost increase
Solution Approach 1:
The patent makes existing sensors perform multiple functions by using them to generate data for multiple algorithm types through translation. A single available sensor can provide data for multiple different algorithm requirements by translating its output into different sensor modalities, eliminating the need for dedicated redundant sensors for each algorithm
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Apparatuses, methods and storage medium associated with compensating for a sensor deficiency in a heterogeneous sensor array are disclosed herein. In embodiments, an apparatus may include a compute device to aggregate perception data from individual perception pipelines, each of which is associated with respective one of different types of sensors of a heterogeneous sensor set, to identify a characteristic associated with a space to be monitored by the heterogeneous sensor set; detect a sensor deficiency associated with a first sensor of the sensors; and in response to a detection of the sensor deficiency, derive next perception data for more than one of the individual perception pipelines from sensor data originating from at least one second sensor of the sensors. Other embodiments may be disclosed or claimed.