CN-117191075-B - Visual inertial odometer system for synchronous time synchronization of multiple sensors and time synchronization method thereof

CN117191075BCN 117191075 BCN117191075 BCN 117191075BCN-117191075-B

Abstract

The application discloses a visual inertial odometer system for synchronous time synchronization of multiple sensors and a time synchronization method thereof, wherein the system comprises a power calculation platform, an MCU, a data bus, a communication interface and more than two sensors; the power calculation platform is connected with the MCU through the data bus, the MCU is connected with the sensor through the communication interface, the power calculation platform is also connected with the communication interface and the sensor respectively, the power calculation platform is used for carrying out initialization parameter setting on the MCU, obtaining sensor data measured by the sensor and carrying out preset processing, the MCU is used for controlling the triggering mode of the sensor according to the initialization parameter, and the sensor is used for obtaining the sensor data. The application can reduce the synchronization time error and improve the compatibility of the sensors, and can be widely applied to the field of synchronization of multiple sensors.

Inventors

ZENG YU
GUO XIAODONG
ZHANG DA
CHEN ZHUORAN

Assignees

广州紫川电子科技有限公司

Dates

Publication Date: 20260505
Application Date: 20230811

Claims (8)

1. The visual inertial odometer system for synchronous time synchronization of multiple sensors is characterized by comprising a force calculation platform, an MCU, a data bus, a communication interface and more than two sensors; The power computing platform is connected with the MCU through the data bus, the MCU is connected with the sensor through the communication interface, and the power computing platform is also connected with the communication interface and the sensor respectively; The power calculation platform is used for carrying out initialization parameter setting on the MCU, acquiring sensor data measured by the sensor and carrying out preset processing; the MCU is used for controlling the triggering mode of the sensor according to the initialization parameter; The sensor is used for acquiring sensor data; The system is used for executing a time setting method of a visual inertial odometer system, and the time setting method comprises the following steps: Receiving initialization parameters sent by a computing platform, wherein the initialization parameters comprise triggering modes of various sensors; carrying out sensor frame data statistics according to the initialization parameters to obtain frame statistics data; Synchronously triggering the corresponding sensor according to the frame statistics data and the triggering mode; Determining trigger frame information of the sensor from the frame statistics data, and transmitting the trigger frame information to the computing platform; the step of synchronously triggering the corresponding sensor according to the frame statistics data and the triggering mode comprises the following steps: determining the output frame rate of each sensor according to the frame statistics; Determining a synchronous frame rate according to the output frame rate of each sensor; Synchronously triggering the corresponding sensor in the synchronous frame rate and the triggering mode; The key frame marking function of the MCU realizes that the power computing platform distinguishes current key frame information, the working principle of the key frame marking comprises that the MCU carries out frame calculation and carries out key frame marking when the condition of triggering the camera is met, and the key frame information is marked on the current inertia measuring unit data, so that the power computing platform is convenient to distinguish.
2. The visual inertial odometer system of claim 1, wherein the computing platform is an ARM processor or a PC host.
3. A multi-sensor synchronous pair visual odometer system of claim 1, the UART is characterized in that the data bus adopts a UART; the communication interface adopts GPIO.
4. A multi-sensor synchronizing time visual odometer system according to any of claims 1-3, wherein the sensors further comprise a range radar and a positioning sensor.
5. A multi-sensor synchrony pair visual odometer system as claimed in claim 4, wherein the cameras comprise monocular and multi-eye cameras; the range radars include TOF and lidar; The inertial measurement unit comprises a gyroscope, an accelerometer and a magnetic field meter; the positioning sensor comprises an RTK sensor, a GPS sensor and a Beidou navigation positioning sensor.
6. A multi-sensor synchronizing time visual odometer system of claim 1, wherein, the initialization parameters comprise an active triggering mode and a passive triggering mode of the sensor, a sensor frame interval triggering parameter, a single piece triggering mode and a multi-condition triggering mode of the sensor; the method for receiving the initialization parameters sent by the computing platform comprises the steps of: And receiving an active triggering mode and a passive triggering mode of the sensor, a sensor frame interval triggering parameter, a single piece triggering mode and a multi-condition triggering mode of the sensor, which are sent by the force calculation platform.
7. The visual odometer system of claim 1, wherein the performing sensor frame data statistics based on the initialization parameters to obtain frame statistics comprises: And counting the triggered times of the sensor through MCU interruption times, comparing the triggered times with the communication ratio of the power computing platform to receive frame data, and counting the frame data of the sensor to obtain frame statistics data.
8. A multi-sensor synchronous pair visual odometer system of claim 1, wherein said determining trigger frame information for said sensor from said frame statistics comprises: And determining a trigger frame for synchronously triggering each sensor according to the synchronous frame rate, and marking the trigger frame by a key frame to obtain trigger frame information.

Description

Visual inertial odometer system for synchronous time synchronization of multiple sensors and time synchronization method thereof Technical Field The application relates to the field of multi-sensor synchronous time synchronization, in particular to a visual inertial odometer system for multi-sensor synchronous time synchronization and a time synchronization method thereof. Background The sensor synchronization time setting method of the conventional visual inertial odometer system comprises software synchronization and hardware synchronization, wherein the software synchronization has the problems of large synchronization time error, data delay and the like, and the hardware synchronization has the problems of limited synchronization frequency, sensor type selection, access quantity and the like. Disclosure of Invention In view of the above, the present application provides a visual inertial odometer system for synchronous synchronization of multiple sensors and a synchronization method thereof, so as to reduce synchronization time errors and improve compatibility of the sensors. An aspect of the present application provides a visual inertial odometer system for synchronizing pairs of multiple sensors, comprising: the system comprises a computing platform, an MCU, a data bus, a communication interface and more than two sensors; The power computing platform is connected with the MCU through the data bus, the MCU is connected with the sensor through the communication interface, and the power computing platform is also connected with the communication interface and the sensor respectively; The power calculation platform is used for carrying out initialization parameter setting on the MCU, acquiring sensor data measured by the sensor and carrying out preset processing; the MCU is used for controlling the triggering mode of the sensor according to the initialization parameter; the sensor is used for acquiring sensor data. Optionally, the computing platform adopts an ARM processor or a PC host. Optionally, the data bus adopts UART; the communication interface adopts GPIO. Optionally, the sensor comprises a camera, a range radar, an inertial measurement unit, and a positioning sensor. Optionally, the camera comprises a monocular camera and a multi-eye camera; the range radars include TOF and lidar; The inertial measurement unit comprises a gyroscope, an accelerometer and a magnetic field meter; the positioning sensor comprises an RTK, a GPS and a Beidou navigation positioning sensor. In another aspect, the present application provides a method for timing a visual inertial odometer system, including: Receiving initialization parameters sent by a computing platform, wherein the initialization parameters comprise triggering modes of various sensors; carrying out sensor frame data statistics according to the initialization parameters to obtain frame statistics data; Synchronously triggering the corresponding sensor according to the frame statistics data and the triggering mode; And determining trigger frame information of the sensor from the frame statistics data, and transmitting the trigger frame information to the computing platform. Optionally, the initialization parameters include an active triggering mode and a passive triggering mode of the sensor, a sensor frame interval triggering parameter, and a single piece triggering mode and a multi-condition triggering mode of the sensor; the method for receiving the initialization parameters sent by the computing platform comprises the steps of: And receiving an active triggering mode and a passive triggering mode of the sensor, a sensor frame interval triggering parameter, a single piece triggering mode and a multi-condition triggering mode of the sensor, which are sent by the force calculation platform. Optionally, the performing sensor frame data statistics according to the initialization parameter to obtain frame statistics data includes: And counting the triggered times of the sensor through MCU interruption times, comparing the triggered times with the communication ratio of the power computing platform to receive frame data, and counting the frame data of the sensor to obtain frame statistics data. Optionally, the step of synchronously triggering the corresponding sensor according to the frame statistics and the triggering mode includes: determining the output frame rate of each sensor according to the frame statistics; Determining a synchronous frame rate according to the output frame rate of each sensor; and synchronously triggering the corresponding sensor in the synchronous frame rate and the triggering mode. Optionally, the determining trigger frame information of the sensor from the frame statistics includes: And determining a trigger frame for synchronously triggering each sensor according to the synchronous frame rate, and marking the trigger frame by a key frame to obtain trigger frame information. The application carries out initialization parameter co