Resistive Memory Sensor Compression With On-Device Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional sensor devices require complex off-device reconstruction algorithms for compressed sensing, which are power-intensive and inefficient, especially in Internet of Things (IoT) systems where energy efficiency is crucial.
Innovation Solution
A sensor device with computational memory and electronic circuitry that performs matrix-vector multiplications using memristive arrays to compress and reconstruct signals on-device, employing an approximate message passing algorithm for efficient signal processing and energy management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If compressed sensing is used to reduce acquisition energy and increase frame rate, then energy efficiency is improved, but complex reconstruction algorithms are required
Solution Approach 1:
The patent merges the sensing function with the reconstruction function into a single integrated sensor device. The computational memory is coupled with electronic circuitry to perform both compression (y=Ax0) and reconstruction (x0=ATy) operations on-device, eliminating the need for separate external processing systems and reducing overall system complexity despite maintaining algorithmic sophistication.
Solution Approach 2:
The sensor device performs self-service by executing the complex reconstruction algorithm internally using its integrated computational memory and electronic circuitry. This allows the device to autonomously recover original signals from compressed measurements without requiring external computational resources, thereby reducing acquisition energy while managing algorithmic complexity through on-device processing.
2Measurement precision
If off-device reconstruction algorithms are used, then signal recovery is achieved, but power consumption increases
Solution Approach 1:
The patent extracts the reconstruction function from external off-device processing and integrates it directly into the sensor device. By placing the computational memory and electronic circuitry within the sensor, the system eliminates the need to transmit compressed data externally for reconstruction, thereby reducing power consumption while maintaining signal recovery accuracy through on-device processing.
Solution Approach 2:
The computational memory acts as an intermediary between the compression stage and the reconstruction stage. It stores the measurement matrix and intermediate computation results, enabling efficient on-device reconstruction without requiring external computational resources, thus reducing power consumption while maintaining measurement precision.
3Productivity
If on-device compression and reconstruction are performed, then external processing is reduced, but device complexity increases
Solution Approach 1:
The patent transitions from a traditional two-stage architecture (separate compression and reconstruction systems) to an integrated on-device processing architecture. By adding the computational memory dimension within the sensor device, the system achieves both compression and reconstruction functions in one location, improving productivity through reduced data transmission while managing complexity through functional integration.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enables energy-efficient signal acquisition and reconstruction within the sensor device, reducing the need for external processing and minimizing power consumption, while maintaining effective signal recovery.
Implementation Method 1
A resistive memory element may be defined as a memory element whose electrical resistance can be changed by applying an electrical programming signal to the resistive memory element
Implementation Method 2
The sensor device with computational memory and electronic circuitry that performs matrix-vector multiplications using memristive arrays to compress and reconstruct signals on-device
Data Source
AI summary
A sensor device comprising a computational memory and electronic circuitry. The sensor device is configured to receive an input signal, to compress the input signal into a compressed signal and to compute a reconstructed signal from the compressed signal. The electronic circuitry is configured to perform a reconstruction algorithm to compute the reconstructed signal. The computational memory is configured to compute the compressed signal and partial results of the reconstruction algorithm. A related method and a related design structure may be provided.