# US 2023/0409668 A1 · Sensor calibration optimization system and method

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2023/0409668 A1
- Document ID: US20230409668A1
- Status: Published application
- Kind: A1
- Application number: 17/845,906
- Filed: 2022-06-21
- Published: 2023-12-21
- Earliest claimed: 2022-06-21
- Assignees: Deepen AI, Inc.
- Applicants: Deepen AI, Inc.
- Inventors: Cheuksan WANG; Naman JAIN; Krishna Chaitanya Majeti
- Web page: https://patentgrep.com/patent/US20230409668A1

## CPC

- G06F 17/17: Function evaluation by approximation methods, e.g. inter- or extrapolation, smoothing, least mean square method ({G06F17/18 takes precedence } ; interpolation for numerical control G05B19/18)
- G06F 17/40: Data acquisition and logging (for input to computer G06F3/00)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

A system and method for calibration optimization. The system includes a plurality of sensors, a processing circuitry and a memory, the memory containing instructions that, when executed by the processing circuitry, configure the system to receive a signal from a plurality of sensors, perform a calibration on the plurality of sensors, the calibration having a predetermined vector size corresponding to a plurality of parameters of the calibrations, and perform an optimization after the calibration. The optimization includes a minimization that includes applying a loop constraint.

## Claims

### Claim 1

A calibration optimization method, comprising:
receiving, by a processor, a signal from a plurality of sensors;
performing, by the processor, a calibration on the plurality of sensors, the calibration having a predetermined vector size corresponding to a plurality of parameters of the calibration;
performing, by the processor, an optimization of the calibration on the plurality of sensors after the calibration, wherein the optimization includes a minimization that includes applying a loop constraint, and minimizing ∥C1′−C1∥2+∥C2′−C2∥2+∥C3′−C3∥2 such that M(C3′)*M(C2′)*M(C1′)=I, wherein:
C1, C2, and C3 each represents a result of the calibration performed on a pair of sensors from the plurality of sensors, and forms a closed loop with each other,
C1′, C2′, C3′ each represents a result of the optimization of the calibration for the respective C1, C2, and C3,
M(C1′), M(C2′), and M(C3′) represent a 4×4 matrix converted from respective calibrations C1′, C2′, and C3′ towards which the optimization is applied, and
I is an identity matrix describing an ideal value; and
minimizing errors of subsequent signals from the plurality of sensors based on the optimization.

### Claim 2 (depends on 1)

The method of claim 1 wherein the calibration includes forming a calibration loop, and the loop constraint includes the formed calibration loop.

### Claim 3 (depends on 1)

The method of claim 1, wherein the calibration includes an extrinsic calibration, and the minimization further comprises subjecting the minimized values to a Levenberg Marquardt algorithm.

### Claim 4 (depends on 1)

The method of claim 1, wherein the calibration includes a temporal calibration, and the minimization comprises:
minimizing (T1′−T1)2+(T2′−T2)2+(T3′−T3)2 such that T1′+T2′+T3′=0, wherein:
T1, T2, and T3 each represents a result of the temporal calibration performed on a pair of sensors from the plurality of sensors, and forms a closed loop with each other, and
T1′, T2′, T3′ each represents a result of the optimization of the temporal calibration for the respective T1, T2, and T3; and
subjecting the minimized values to a Levenberg Marquardt algorithm.

### Claim 5 (depends on 4)

The method of claim 4, wherein the temporal calibration further includes measuring a time difference in timestamp amongst the plurality of sensors.

### Claim 6 (depends on 1)

The method of claim 1, further comprising estimating an error in the calibration based on a calibration loop residual error, wherein the calibration loop residual error is a deviation of the calibration from the loop constraint.

### Claim 7 (depends on 6)

The method of claim 6, wherein the estimated error in the calibration is determined based on one of a size of the loop constraint, or a vector of a plurality of parameters.

### Claim 8

A non-transitory computer readable medium having stored thereon instructions for causing a processing circuitry to execute a process, the process comprising:
receiving, by the processing circuitry, a signal from a plurality of sensors selected by a user;
performing, by the processing circuitry, a calibration on the plurality of sensors;
forming, by the processing circuitry, a calibration loop based on the plurality of sensors;
performing, by the processing circuitry, an optimization of the calibration on the signals from the plurality of sensors after the calibration is performed, including a minimization comprising minimizing ∥C1′−C1∥2+∥C2′−C2∥2+∥C3′−C3∥2 such that M(C3′)*M(C2′)*M(C1′)=I, wherein:
C1, C2, and C3 each represents a result of the calibration performed on a pair of sensors from the plurality of sensors, and forms a closed loop with each other,
C1′, C2′, C3′ each represents a result of the optimization of the calibration for the respective C1, C2, and C3,
M(C1′), M(C2′), and M(C3′) represent a 4×4 matrix converted from respective calibrations C1′, C2′, and C3′ towards which the optimization is applied, and
I is an identity matrix describing an ideal value;
determining, by the processing circuitry, an optimized calibration value by applying a loop constraint based on the formed calibration loop; and
applying the optimized calibration value to the plurality of sensors to minimize errors of subsequent signals from the plurality of sensors.

### Claim 9 (depends on 8)

The non-transitory computer readable medium of claim 8, wherein the calibration includes an extrinsic calibration, and
the minimization further comprises subjecting the minimized values to a Levenberg Marquardt algorithm.

### Claim 10 (depends on 8)

The non-transitory computer readable medium of claim 8, wherein the calibration includes a temporal calibration, and
the minimization comprises:
minimizing (T1′−T1)2+(T2′−T2)2+(T3′−T3)2 such that T1′+T2′+T3′=0, wherein:
T1, T2, and T3 each represents a result of the temporal calibration performed on a pair of sensors from the plurality of sensors, and forms a closed loop with each other, and
T1′, T2′, T3′ each represents a result of the optimization of the temporal calibration for the respective T1, T2, and T3; and
subjecting the minimized values to a Levenberg Marquardt algorithm.

### Claim 11 (depends on 10)

The non-transitory computer readable medium of claim 10, wherein the temporal calibration further includes measuring a time difference in timestamp amongst the plurality of sensors.

### Claim 12 (depends on 8)

The non-transitory computer readable medium of claim 8, further comprising estimating an error in the calibration based on a calibration loop residual error, wherein the calibration loop residual error is a deviation of the calibration from the loop constraint.

### Claim 13 (depends on 12)

The non-transitory computer readable medium of claim 12, wherein the estimated error in the calibration is determined based on one of a size of the loop constraint, or a vector of a plurality of parameters.

### Claim 14

A sensing system, comprising:
a plurality of sensors;
a processing circuitry; and
a memory, the memory containing instructions that, when executed by the processing circuitry, configure the sensing system to:
receive, by the processing circuitry, a signal from the plurality of sensors;
perform, by the processing circuitry, a calibration on the plurality of sensors, the calibration having a predetermined vector size corresponding to a plurality of parameters of the calibration;
perform, by the processing circuitry, an optimization of the calibration on the plurality of sensors after the calibration, wherein the optimization includes a minimization that includes applying a loop constraint, wherein:
the calibration includes an extrinsic calibration, and
the optimization includes a minimization comprising:
minimizing ∥C1′−C1∥2+∥C2′−C2∥2+∥C3′−C3∥2 such that M(C3′)*M(C2′)*M(C1′)=I, wherein:
C1, C2, and C3 each represents a result of the calibration performed on a pair of sensors from the plurality of sensors, and forms a closed loop with each other,
C1′, C2′, C3′ each represents a result of the optimization of the calibration for the respective C1, C2, and C3,
M(C1′), M(C2′), and M(C3′) represent a 4×4 matrix converted from respective calibrations C1′, C2′, and C3′ towards which the optimization is applied, and
I is an identity matrix describing an ideal value,
minimizing errors of subsequent signals from the plurality of sensors based on the optimization, and
subjecting the minimized values to a Levenberg Marquardt algorithm.

### Claim 15 (depends on 14)

The sensing system of claim 14, wherein the calibration includes a temporal calibration, and the minimization comprises:
minimizing (T1′−T1)2+(T2′−T2)2+(T3′−T3)2 such that T1′+T2′+T3′=0, wherein:
T1, T2, and T3 each represents a result of the temporal calibration performed on a pair of sensors from the plurality of sensors, and forms a closed loop with each other, and
T1′, T2′, T3′ each represents a result of the optimization of the temporal calibration for the respective T1, T2, and T3; and
subjecting the minimized values to a Levenberg Marquardt algorithm.

### Claim 16 (depends on 15)

The sensing system of claim 15, wherein the temporal calibration further includes
measuring a time difference in timestamp amongst the plurality of sensors.

### Claim 17 (depends on 14)

The sensing system of claim 14, wherein the sensing system is further configured to estimating an error in the calibration based on a calibration loop residual error, wherein the calibration loop residual error is a deviation of the calibration from the loop constraint.

### Claim 18 (depends on 17)

The sensing system of claim 17, wherein the estimated error in the calibration is determined based on one of a size of the loop constraint, or a vector of a plurality of parameters.
