Sensor Unique ID via Bad Pixel Pattern Analysis

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional methods for generating hardware identification values for sensors are either non-unique, increasing system complexity and cost, or rely on external cryptography chips, which are costly and complex.

Innovation Solution

A method to generate a unique hardware identification value based on the fabrication-dependent random characteristics of a sensor, specifically using the pattern of stable bad pixels in an image sensor, which is virtually unclonable and does not require additional hardware, thereby reducing complexity and cost.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If hardware identification values are derived from configurable media access control addresses, then the system complexity is reduced, but the hardware uniqueness is not guaranteed as multiple copies can be configured with the same address

Engineering Contradiction:
Improvesystem complexityVSAvoidhardware uniqueness
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The sensor itself generates its unique identification value through its inherent fabrication-dependent characteristics (bad pixels), without requiring external cryptography chips or configurable addresses. The sensor serves itself to create a unique identifier that is both simple to implement and guaranteed to be unique to each physical sensor instance.

Inventive Principle:
Principle #25Self-service

2Reliability

If external cryptography chips are used to generate hardware identification values, then the hardware uniqueness is improved, but the system complexity and cost increase

Engineering Contradiction:
Improvehardware uniquenessVSAvoidsystem complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The invention extracts the unique identification capability directly from the sensor's inherent physical characteristics (specifically, the pattern of bad pixels resulting from fabrication variations), removing the need for external cryptography chips. This extraction approach maintains hardware uniqueness while eliminating the added complexity and cost of separate cryptographic hardware.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

Instead of using physical cryptography chips, the invention creates a digital copy or representation of the sensor's unique physical characteristics (the bad pixel pattern) as its identification value. This digital representation captures the essence of hardware uniqueness without requiring physical cryptographic components.

Inventive Principle:
Principle #26Copying

3Reliability

If external cryptography chips are used to generate hardware identification values, then the hardware uniqueness is improved, but the cost increases

Engineering Contradiction:
Improvehardware uniquenessVSAvoidcost
Core Design Contradiction:
ReliabilityVSEase of manufacture

Solution Approach 1:

The invention uses the sensor's inherent, inexpensive fabrication variations (bad pixels) as the source of uniqueness, replacing expensive cryptography chips with a feature that already exists in every sensor due to normal manufacturing processes. This approach makes unique identification economically viable for mass-produced sensors.

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

Data Source

PatentUS10742915B2Unique identification value for a sensor
Publication Date: 2020.08.11 AMBARELLA INT LP
  • US10742915B2 patent drawing
  • US10742915B2 patent drawing
  • US10742915B2 patent drawing

AI summary

An apparatus includes an interface and a processor. The interface may be configured to receive a plurality of sensed values from a sensor under a plurality of conditions. The processor may be connected to the interface and may be configured to (i) generate a plurality of patterns in response to the sensed values, where each of the patterns includes a plurality of positions of a plurality of bad values among the sensed values in a corresponding one of the conditions, (ii) identify a plurality of consistently bad values among the bad values in the patterns and (iii) indicate that the sensor has been replaced where the positions associated with the consistently bad values do not correlate to a plurality of known positions of the consistently bad values previously identified in the sensor.