# US 2022/0019434 A1 · System and method for pipelined time-domain computing using time-domain flip-flops and its application in time-series analysis

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

- Number: US 2022/0019434 A1
- Document ID: US20220019434A1
- Status: Published application
- Kind: A1
- Application number: 17/311,849
- Filed: 2019-12-18
- Published: 2022-01-20
- Earliest claimed: 2018-12-18
- Assignees: Northwestern University
- Applicants: Northwestern University
- Inventors: Jie Gu; Zhengyu Chen
- Web page: https://patentgrep.com/patent/US20220019434A1

## CPC

- G06F 9/30029: {Logical and Boolean instructions, e.g. XOR, NOT}
- G06T 3/18: Image warping, e.g. rearranging pixels individually
- G06F 7/605: {Additive or subtractive mixing of two pulse rates into one (beat-frequency oscillators H03B21/00; input circuits of electric counters, e.g. up-down counters H03K21/00)}
- G06F 9/3001: {Arithmetic instructions}
- G06F 9/3555: {using scaling, e.g. multiplication of index}
- G06F 17/18: for evaluating statistical data {, e.g. average values, frequency distributions, probability functions, regression analysis (forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04)}
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

Systems and/or methods can include a ring based inverter chain that constructs multi-bit flip-flops that store time. The time flip-flops serve as storage units and enable pipeline operations. Single cells used in time series analysis, such as dynamic time warping are rendered by the time-domain circuits. The circuits include time flip-flops, Min, and ABS circuits. A and the matrix can be constructed through the single cells.

## Claims

### Claim 1

A time-domain hardware implemented method for processing time sequences signals, comprising:
performing, using a scalable pipelined structure that comprises multi-bit ring-based time flip-flops (TFF), dynamic time warping (DTW) between two time sequences A and B to detect data similarities as a node Di,j in the pipeline structure, wherein the time sequences A and B each comprises time sample signals Ai and Bj of variable speed, respectively, wherein:
each of the Ai and Bj time sample signals has a numerical value which is proportional to a duration of time pulse at a time instance i and j, respectively, and
the node Di,j is stored as a resultant time pulse T(Di,j) in the TFF for a next clock cycle operation, and the resultant time pulse T(Di,j) is obtained by directly performing at least a time-domain pulses comparison operation on a pair of time sample signals Ai and Bj.

### Claim 2 (depends on 1)

The time-domain hardware implemented method of claim 1, wherein the time-domain pulses comparison operation comprising performing an absolute difference operation (ABS) on the pair of time sample signals Ai and Bj to obtain the resultant time pulse T(Di,j), wherein T(Di,j)=ABS (Ai−Bj) and the resultant time pulse T(Di,j) is stored in the TFF for carrying out more time-domain pulses comparison operation with neighboring nodes in a next clock cycle to generate new resultant time pulses T(Di,j), until all nodes Di,j within an i×j dimension matrix between the time sequences A and B are determined.

### Claim 3 (depends on 2)

The time-domain hardware implemented method of claim 2, wherein a path of nodes Di,j within the i×j dimension matrix having the lowest pulse widths forms a warping path which aligns respective pairs of time sample signals Ai and Bj between the time sequences A and B, and wherein a pulse width value of a terminating node D*i,j at an opposite diagonal corner from a starting node of an i×j dimension matrix represents a final distance which indicates how much similarities between the time sequences A and B.

### Claim 4 (depends on 1)

The time-domain hardware implemented method of claim 1, wherein the time-domain pulses comparison operation comprising taking a minimum value operation (MIN) of three ancestor nodes (Di-1,j, Di-1,j-1, Di,j-1), wherein T(Di,j)=MIN (Di-1,j, Di-1,j-1, Di,j-1) and the pulse T(Di,j) forms a node Di,j to populate an i×j dimension matrix between the time sequences A and B.

### Claim 5 (depends on 4)

The time-domain hardware implemented method of claim 4, wherein the time-domain pulses comparison operation comprising operating in a pipeline mode by reading the resultant time pulse T(Di,j) of an absolute difference operation (ABS) operation stored in the TFF, for a next cycle MIN operation to form respective nodes Di,j to populate the i×j dimension matrix.

### Claim 6 (depends on 5)

The time-domain hardware implemented method of claim 5, comprising using a multiplexer (MUX) to by-pass node pulse data reading from an input TFF to speed up the DTW operations.

### Claim 7 (depends on 1)

The time-domain hardware implemented method of claim 1, wherein the time signal samples Ai and Bj are read in digital from the sequences A and B, and the time signal samples Ai and Bj are converted into quantized pulse widths prior to performing an absolute difference operation ABS or a minimum value operation (MIN) operation.

### Claim 8 (depends on 5)

The time-domain hardware implemented method of claim 5, further comprising fine tuning diagonally in the i×j matrix, the stored pulse widths values of the nodes Di,j.

### Claim 9 (depends on 2)

The time-domain hardware implemented method of claim 2, comprising using a NAND gate having a plurality of NAND gate inputs to perform the ABS operation, wherein only one of the plurality of NAND gate inputs is buffered by an inverter.

### Claim 10 (depends on 4)

The time-domain hardware implemented method of claim 4, comprising using a multi-input NOR gate to directly perform the MIN operation on time-domain pulses.

### Claim 11 (depends on 1)

The time-domain hardware implemented method of claim 1, wherein the TFF in the pipelines are used to store the resultant time pulse T(Di,j) for a future clock cycle operations and to perform accumulation operations in DTW calculations when enabled.

### Claim 12 (depends on 1)

The time-domain hardware implemented method of claim 1, wherein the pipelined structure is scalable by simply cascading the pipelined structure horizontally and vertically to increase sequence length for reading more time sample signals Ai and Bj simultaneously.
