CN-116105737-B - RISC-V-based pedestrian autonomous positioning method

CN116105737BCN 116105737 BCN116105737 BCN 116105737BCN-116105737-B

Abstract

The invention discloses a pedestrian autonomous positioning method based on RISC-V, which comprises the following steps of obtaining three-dimensional data of pedestrian motion in an acceleration sensor and a gyroscope sensor, calculating an attitude angle, converting the attitude angle from a carrier coordinate system to a coordinate transformation matrix of a navigation coordinate system, carrying out inertial navigation basic solution according to the three-dimensional data of pedestrian motion, the attitude angle and the coordinate transformation matrix to solve the current attitude, speed and position of the pedestrian, carrying out zero-speed detection based on a neural network, judging whether the moment point of the current state of the pedestrian is a zero-speed point, if so, entering the next step, otherwise, returning to the first step, carrying out positioning solution by using Kalman filtering EKF to update the current attitude, position and speed of the pedestrian, and completing the autonomous positioning of the pedestrian according to the updated attitude, the updated position and the updated speed. The invention can reduce noise error, reduce calculation amount, reduce time and resource required by execution of the processor, and has good portability.

Inventors

YAN BO
SHI CHAOFAN
WANG PENGFEI
ZHANG ZIHENG
DENG YUZHI

Assignees

电子科技大学

Dates

Publication Date: 20260505
Application Date: 20230213

Claims (3)

1. The pedestrian autonomous positioning method based on RISC-V is characterized by comprising the following steps: s1, acquiring three-dimensional data of pedestrian motion in an acceleration sensor and a gyroscope sensor; S2, calculating an attitude angle according to the three-dimensional data of the pedestrian motion, and converting the attitude angle from a carrier coordinate system to a coordinate transformation matrix of a navigation coordinate system; S3, carrying out inertial navigation basic calculation according to the three-dimensional data of the pedestrian motion, the attitude angle and the coordinate transformation matrix to solve the current attitude, the speed and the position of the pedestrian; s4, detecting zero speed based on a neural network through the current gesture, speed and position of the pedestrian, judging whether the moment point of the current state of the pedestrian is the zero speed point, if so, entering a step S5, otherwise, entering a step S1; s5, carrying out positioning calculation to update the current posture, position and speed of the pedestrian by using Kalman filtering EKF, and obtaining the updated posture, the updated position and the updated speed; S6, completing autonomous positioning of pedestrians according to the updated posture, the updated position and the updated speed; The specific implementation manner of the step S2 is as follows: S2-1, calculating a posture angle according to the three-dimensional data of the pedestrian motion, wherein the posture angle comprises a roll angle Pitch angle And heading angle ; S2-2, roll angle Pitch angle And heading angle A 3*3 coordinate transformation matrix which is converted from the carrier coordinate system to the navigation coordinate system, and the values of the coordinate transformation matrix are stored in a 3*3 floating point array C; the specific implementation manner of the step S3 is as follows: s3-1, respectively averaging the three-dimensional data of pedestrian movement collected by the gyroscope sensor to obtain And average value Storing array variable gyr 3 of floating point type in sequence, wherein t is time; is the average value of the first dimension data; is the average value of the second dimension data; Is the third dimension data average; s3-2 according to Computing a skew symmetry matrix ; S3-3, according to the formula: Obtaining an updated pose matrix As the current posture of the pedestrian, wherein, 3*3 Floating point array C; Is a unit matrix; Is a sampling time interval; is an oblique symmetry matrix of the gyroscope at the moment t; S3-4, according to the updated gesture matrix And obtaining the current speed and the current position of the pedestrian.
2. The pedestrian autonomous positioning method based on RISC-V as set forth in claim 1, wherein the specific implementation manner of step S3-4 is as follows: according to the formula: Obtaining inertial navigation acceleration over time Speed of And position Wherein, the method comprises the steps of, A matrix formed by the average value of the pedestrian motion three-dimensional data collected by the acceleration sensor; is the transpose of the matrix; is the speed of the last moment; Is the position of the last moment; Is the acceleration at the previous moment.
3. The pedestrian autonomous positioning method based on RISC-V as set forth in claim 2, wherein the specific implementation manner of step S5 is as follows: s5-1, according to the formula: Obtaining covariance matrix of predictive vector at k moment Wherein Q represents a covariance matrix of the noise vector; a state transition matrix is represented and is used to represent, , The unit matrix is 3*3 of the unit matrix, For a diagonally symmetrical cross matrix of triaxial acceleration measurements in a navigational coordinate system, Represents the acceleration at time k in the z-axis direction in the navigation coordinate system, Represents the acceleration at time k in the y-axis direction in the navigation coordinate system, The acceleration at time k in the x-axis direction in the navigation coordinate system is represented by (-) T , which represents the transpose of the matrix; s5-2, according to the formula: obtaining error vector Wherein, the method comprises the steps of, Is an observation matrix; r represents a covariance matrix of a multiple independent normal distribution, (. Cndot.) -1 represents the inverse of the matrix; and S5-3, updating the posture, the position and the speed of the pedestrian according to the error vector to obtain the updated posture, the updated position and the updated speed.

Description

RISC-V-based pedestrian autonomous positioning method Technical Field The invention relates to the technical field of vector processors, in particular to a pedestrian autonomous positioning method based on RISC-V. Background RISC-V is an open source instruction set architecture based on the reduced instruction set principle. The RISC-V adopts a modularized design mode, and comprises other expansion instruction sets such as RV32M, RV32F, RV32V and the like besides the basic RV 32I. RV32V is a vector extended instruction set of RISC-V, unlike conventional SIMD instructions, which supports variable length vector registers while separating the length of the vector from the maximum number of operands that can be performed per clock cycle. Based on the RV32V vector expansion instruction set, the performance of a plurality of algorithms can be improved. Positioning navigation algorithms are widely used in various fields of today's society, wherein inertial navigation systems based on inertial sensors can realize autonomous positioning navigation without relying on external information. The conventional inertial navigation system has high requirements on the sensor, and is not applied to pedestrian positioning. The existing micromechanical inertial measurement unit has the advantages of small volume, weight, price and the like, and is widely applied to the pedestrian inertial navigation system. The basic inertial measurement unit is typically composed of a three-axis accelerometer and a three-axis gyroscope. However, current measurement methods suffer from noise interference, and many positioning navigation algorithms use kalman filtering to estimate the state of the dynamic system to reduce noise errors. However, since the kalman filtering algorithm contains a large amount of operations, the execution of the algorithm in the processor with the traditional architecture needs a large amount of resources and time, which also limits the application of the positioning navigation algorithm in the equipment with the shortage of resources and limited performance. Disclosure of Invention Aiming at the defects in the prior art, the pedestrian autonomous positioning method based on RISC-V solves the problems that the prior art is easy to be interfered by noise, has poor portability and alteration, large resource and time expenditure and limited application in equipment with tense resource and limited performance. In order to achieve the aim of the invention, the technical scheme adopted by the invention is that the pedestrian autonomous positioning method based on RISC-V comprises the following steps: s1, acquiring three-dimensional data of pedestrian motion in an acceleration sensor and a gyroscope sensor; S2, calculating an attitude angle according to the three-dimensional data of the pedestrian motion, and converting the attitude angle from a carrier coordinate system to a coordinate transformation matrix of a navigation coordinate system; S3, carrying out inertial navigation basic calculation according to the three-dimensional data of the pedestrian motion, the attitude angle and the coordinate transformation matrix to solve the current attitude, the speed and the position of the pedestrian; s4, detecting zero speed based on a neural network through the current gesture, speed and position of the pedestrian, judging whether the moment point of the current state of the pedestrian is the zero speed point, if so, entering a step S5, otherwise, entering a step S1; s5, carrying out positioning calculation to update the current posture, position and speed of the pedestrian by using Kalman filtering EKF, and obtaining the updated posture, the updated position and the updated speed; S6, completing autonomous positioning of pedestrians according to the updated posture, the updated position and the updated speed. Further, the specific implementation manner of step S2 is as follows: s2-1, calculating a posture angle according to three-dimensional data of pedestrian motion, wherein the posture angle comprises a roll angle gamma, a pitch angle theta and a course angle phi; S2-2, converting the roll angle gamma, the pitch angle theta and the heading angle phi from a carrier coordinate system to a 3*3 coordinate transformation matrix of a navigation coordinate system, and storing the values of the coordinate transformation matrix into a 3*3 floating point array C. Further, the specific implementation manner of step S3 is as follows: S3-1, respectively averaging pedestrian motion three-dimensional data collected by a gyroscope sensor to obtain omega x(t)、ωy(t)、ωz (t), and respectively storing the average value omega x(t)、ωy(t)、ωz (t) into an array variable gyr 3 of a floating point type in sequence, wherein t is time, omega x (t) is a first dimension data average value, omega y (t) is a second dimension data average value, and omega z (t) is a third dimension data average value; S3-2, calculating an oblique symmetry matrix omega t accord