AI Sensor Estimation Models for Low-Complexity Terminals

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing technologies face challenges in efficiently processing and estimating sensor information using artificial intelligence, particularly in scenarios where multiple types of sensor data need to be processed on a single terminal, due to high computational loads and the cost of equipping devices with specialized sensors.

Innovation Solution

An information processing apparatus and method that collects sensor information from multiple sensors, generates learned models to estimate second sensor information based on first sensor information, and provides these models as services to terminals, enabling the estimation of sensor information without requiring all sensors to be present on the terminal.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple specialized sensors are equipped on each terminal device, then sensor information coverage and measurement precision are improved, but device cost and complexity increase significantly

Engineering Contradiction:

Improvesensor information accuracyVSAvoiddevice structure complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent creates virtual copies of specialized sensor data through AI-generated learned models. Instead of physically equipping every terminal with multiple specialized sensors, the system generates synthetic sensor information that replicates what those specialized sensors would detect, based on data from commonly available sensors. This allows terminals to access specialized sensor data without physically having those sensors.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent makes common sensors serve multiple functions by using their data to generate information for multiple different sensor types through AI processing. A single common sensor (like a camera) can generate data that serves as input for multiple different learned models, effectively making it perform the function of multiple specialized sensors simultaneously.

Inventive Principle:

Principle #6Universality (Multi-functionality)

2Adaptability or versatility

If multiple specialized sensors are equipped on each terminal device, then sensor information coverage is improved, but manufacturing cost increases

Engineering Contradiction:

Improvesensor coverageVSAvoiddevice manufacturing cost

Core Design Contradiction:

Adaptability or versatilityVSEase of manufacture

Solution Approach 1:

Inventive Principle:

Principle #26Copying

3Measurement precision

If learned models are generated and stored for every possible sensor combination, then service accuracy is improved, but memory requirements and system complexity increase

Engineering Contradiction:

Improveestimation accuracyVSAvoidmodel storage volume

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments the learned model system into two distinct parts: commonly used learned models are pre-generated and stored in the system, while specialized or rare sensor combinations generate learned models on-demand. This segmentation allows the system to maintain high accuracy for common cases while avoiding the need to store excessive models for all possible combinations, reducing overall storage requirements.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary generation and storage of learned models for commonly occurring sensor combinations before they are actually needed. This allows the system to have ready-to-use models for frequent cases, improving response time and accuracy for common scenarios while avoiding the need to generate models repeatedly.

Inventive Principle:

Principle #10Preliminary action

4Speed

If AI processing is performed locally on terminal devices, then service speed is improved, but computational load and energy consumption increase

Engineering Contradiction:

Improveprocessing speedVSAvoidterminal energy consumption

Core Design Contradiction:

SpeedVSUse of energy by moving object

Solution Approach 1:

The patent extracts the computationally intensive AI processing functions from terminal devices and relocates them to a centralized server system. The terminal's role is reduced to collecting sensor data and receiving processed results, while the heavy computational workload of generating and managing learned models is performed remotely on the server, significantly reducing terminal energy consumption.

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS20220383055A1Information processing apparatus and information processing method

Publication Date: 2022.12.01 SONY GROUP CORP

AI summary

Provided is an information processing apparatus that performs processing of sensor information using artificial intelligence. The information processing apparatus includes: a collection unit that collects first sensor information detected by a first sensor and second sensor information detected by a second sensor; a model generation unit that generates a learned model for estimating second sensor information corresponding to first sensor information on the basis of the first sensor information and the second sensor information that have been collected; an accumulation unit that accumulates the learned model; and a providing unit that provides a result of a service based on the learned model. The providing unit provides the learned model to the second apparatus in response to a request from the second apparatus.