WO2023071615A1 - 定位方法、装置、终端和存储介质 - Google Patents

定位方法、装置、终端和存储介质 Download PDF

Info

Publication number
WO2023071615A1
WO2023071615A1 PCT/CN2022/119859 CN2022119859W WO2023071615A1 WO 2023071615 A1 WO2023071615 A1 WO 2023071615A1 CN 2022119859 W CN2022119859 W CN 2022119859W WO 2023071615 A1 WO2023071615 A1 WO 2023071615A1
Authority
WO
WIPO (PCT)
Prior art keywords
measurement data
inertial measurement
speed
terminal
user equipment
Prior art date
Application number
PCT/CN2022/119859
Other languages
English (en)
French (fr)
Inventor
裴璇
Original Assignee
上海瑾盛通信科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 上海瑾盛通信科技有限公司 filed Critical 上海瑾盛通信科技有限公司
Publication of WO2023071615A1 publication Critical patent/WO2023071615A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W64/00—Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • H04W64/006—Locating users or terminals or network equipment for network management purposes, e.g. mobility management with additional information processing, e.g. for direction or speed determination
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/02—Services making use of location information
    • H04W4/025—Services making use of location information using location based information parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/30—Services specially adapted for particular environments, situations or purposes
    • H04W4/33—Services specially adapted for particular environments, situations or purposes for indoor environments, e.g. buildings

Definitions

  • the present application relates to the technical field of positioning, and in particular to a positioning method, device, terminal and storage medium.
  • the inertial navigation data is collected through the inertial measurement unit (IMU) installed in the smart terminal, and the position information or motion trajectory of the smart terminal is determined according to the inertial navigation data.
  • IMU inertial measurement unit
  • the positioning accuracy of the above positioning methods cannot meet the requirements of indoor positioning and navigation applications.
  • an embodiment of the present disclosure provides a positioning method, which includes:
  • the speed prediction network is obtained by training the sample inertial measurement data collected by multiple user devices;
  • the positioning data of the terminal to be tested is determined according to the speed of the terminal to be tested.
  • the acquisition of the inertial measurement data collected by at least one user equipment includes:
  • Attitude transformation and coordinate transformation are performed on the original inertial measurement data to obtain the inertial measurement data.
  • the attitude transformation and coordinate transformation are performed on the original inertial measurement data to obtain the inertial measurement data, including:
  • the conversion of the original inertial measurement data into the first inertial measurement data of the user equipment under the standard attitude includes:
  • the original inertial measurement data is mapped to the coordinate system of the user equipment under the standard attitude to obtain the first inertial measurement data.
  • the above-mentioned conversion of the first inertial measurement data to the world coordinate system to obtain the inertial measurement data includes:
  • the second inertial measurement data is converted to the world coordinate system to obtain the inertial measurement data.
  • the inertial measurement data collected by at least one user equipment includes inertial measurement data collected by multiple user equipments, and the method further includes:
  • the determination of the positioning data of the terminal under test according to the speed of the terminal under test includes:
  • the method also includes:
  • service information is obtained and output; wherein, the service information includes at least one of path guidance information and item introduction information.
  • the training process of the above-mentioned speed prediction network includes:
  • the training sample set includes a plurality of sample inertial measurement data and labels corresponding to each sample inertial measurement data, labeled as actual speed;
  • the initial prediction network is determined as the speed prediction network.
  • the above-mentioned acquisition of the training sample set includes:
  • a training sample set is obtained according to the plurality of sample inertial measurement data and the actual velocity corresponding to each sample inertial measurement data.
  • At least one user equipment includes at least one of a smart phone, a headset, a smart bracelet, a smart watch, and smart glasses.
  • an embodiment of the present disclosure provides a positioning device, which includes:
  • the speed prediction module is used to input the inertial measurement data into the speed prediction network for noise reduction and speed prediction to obtain the speed of the terminal to be tested; the speed prediction network is obtained by training the sample inertial measurement data collected by multiple user devices;
  • the positioning data determination module is used to determine the positioning data of the terminal to be tested according to the speed of the terminal to be tested.
  • the above-mentioned inertial data acquisition module includes:
  • a raw data acquisition submodule configured to acquire raw inertial measurement data collected by at least one user equipment
  • the conversion sub-module is used to perform attitude conversion and coordinate conversion on the original inertial measurement data to obtain the inertial measurement data.
  • the above conversion sub-module is specifically used to convert the original inertial measurement data into the first inertial measurement data of the user equipment under the standard attitude; convert the first inertial measurement data into the world coordinate system to obtain the inertial Measurement data.
  • the above conversion sub-module is specifically used to obtain the attitude angle of the user equipment according to the original inertial measurement data; according to the attitude angle of the user equipment, map the original inertial measurement data to the coordinate system of the user equipment under the standard attitude , get the first inertial measurement data.
  • the above conversion sub-module is specifically used to convert the first inertial measurement data to the coordinate system of the terminal under test according to the corresponding relationship between the coordinate system of the user equipment under the standard attitude and the coordinate system of the terminal under test.
  • the second inertial measurement data is obtained; according to the corresponding relationship between the coordinate system of the terminal to be tested and the world coordinate system, the second inertial measurement data is converted into the world coordinate system to obtain the inertial measurement data.
  • the inertial measurement data collected by at least one user equipment includes inertial measurement data collected by multiple user equipments, and the apparatus further includes:
  • the alignment module is used to perform time alignment on the inertial measurement data collected by each user equipment according to the time stamps of the inertial measurement data collected by each user equipment, and obtain the aligned inertial measurement data;
  • the above-mentioned speed prediction module is specifically used to input the aligned inertial measurement data into the speed prediction network for noise reduction and speed prediction to obtain the speed of the terminal to be tested.
  • the positioning data determination module is used to obtain the trajectory information of the terminal to be tested according to the speed of the terminal to be tested; match the trajectory information with the preset scene map to obtain and output the location information of the terminal to be tested.
  • the device also includes:
  • the service information output module is used to obtain and output service information according to the location information of the terminal to be tested and the preset scene map; wherein, the service information includes at least one of path guidance information and item introduction information.
  • the device also includes:
  • the sample acquisition module is used to obtain a training sample set; wherein, the training sample set includes a plurality of sample inertial measurement data and labels corresponding to each sample inertial measurement data, which are marked as actual speed;
  • the training module is used to train the neural network model based on the training sample set to obtain the initial prediction network
  • the test module is used to test the initial prediction network based on the training sample set, and determine a plurality of test tracks according to the prediction speed obtained by the test;
  • the network determination module is configured to determine the initial prediction network as the speed prediction network when the error between the multiple test tracks is less than a preset error value.
  • the above-mentioned sample acquisition module is specifically used to respectively acquire a plurality of original inertial measurement data collected by the data acquisition device and the actual trajectory collected by the annotation acquisition device; respectively perform attitude conversion and coordinate conversion on each original inertial measurement data , to obtain the sample inertial measurement data corresponding to each original inertial measurement data; after aligning the collection time of multiple sample inertial measurement data with the collection time of the actual trajectory, determine the actual velocity corresponding to each sample inertial measurement data; according to multiple The sample inertial measurement data and the actual velocity corresponding to each sample inertial measurement data obtain a training sample set.
  • At least one user equipment includes at least one of a smart phone, a headset, a smart bracelet, a smart watch, and smart glasses.
  • the speed prediction network is obtained by training the sample inertial measurement data collected by multiple user devices;
  • the positioning data of the terminal to be tested is determined according to the speed of the terminal to be tested.
  • an embodiment of the present disclosure provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
  • the speed prediction network is obtained by training the sample inertial measurement data collected by multiple user devices;
  • the positioning data of the terminal to be tested is determined according to the speed of the terminal to be tested.
  • the above positioning method, device, terminal and storage medium obtain inertial measurement data collected by at least one user equipment, input the inertial measurement data into the speed prediction network for noise reduction and speed prediction, and obtain the speed of the terminal to be tested; according to the terminal to be tested
  • the speed determines the positioning data of the terminal to be tested.
  • the speed of the terminal to be tested is not obtained by the integral method, but the speed of the terminal to be tested is obtained by using the speed prediction network.
  • the speed prediction network is used Noise reduction can be achieved, thereby improving the accuracy of the speed.
  • the position information or trajectory information determined according to the speed will be more accurate, that is, the positioning accuracy is improved, and it can meet the needs of indoor positioning and navigation applications.
  • Fig. 1 is an application environment diagram of a positioning method in an embodiment
  • Fig. 2 is a schematic flow chart of a positioning method in an embodiment
  • Fig. 3 is a schematic flowchart of the step of acquiring inertial measurement data collected by at least one user equipment in an embodiment
  • Fig. 4 is a schematic flow chart of the steps of attitude transformation and coordinate transformation on the original inertial measurement data in one embodiment
  • FIG. 5 is a schematic flow chart of a positioning method in another embodiment
  • FIG. 6 is a schematic flowchart of the steps of determining the positioning data of the terminal to be tested according to the speed of the terminal to be tested in an embodiment
  • Fig. 7 is a schematic flow chart of the step of outputting service information in an embodiment
  • FIG. 8 is a schematic flow diagram of the training steps of the speed prediction network in one embodiment
  • FIG. 9 is a schematic flow chart of the step of obtaining a training sample set in an embodiment
  • Fig. 11 is the second structural block diagram of the positioning device in an embodiment
  • Fig. 12 is the third structural block diagram of the positioning device in an embodiment
  • Fig. 13 is the fourth structural block diagram of the positioning device in an embodiment
  • Fig. 14 is an internal structure diagram of a terminal in an embodiment.
  • the inertial navigation data is collected through the inertial measurement unit (IMU) installed in the smart terminal, and the position information or motion trajectory of the smart terminal is determined according to the inertial navigation data.
  • IMU inertial measurement unit
  • the double integral method is often used to determine the location information or motion trajectory of the smart terminal based on the inertial navigation data. This method is prone to trajectory errors and affects the positioning accuracy, making it difficult for the positioning accuracy to meet the needs of indoor positioning and navigation applications.
  • the inertial measurement data collected by at least one user equipment is obtained first; then the inertial measurement data is input into the speed prediction network for noise reduction and speed prediction, and the speed of the terminal to be tested is obtained; finally, according to the The speed of the terminal under test determines the positioning data of the terminal under test.
  • the speed prediction network is used to obtain the speed of the terminal to be tested, and then to obtain the positioning data of the terminal to be tested, which can reduce trajectory errors, improve positioning accuracy, and meet the needs of indoor positioning and navigation applications.
  • the positioning method provided in this application can be applied to the application environment shown in FIG. 1 .
  • the application environment includes a user equipment 102 and a terminal 104 to be tested.
  • the user equipment 102 can communicate with the terminal under test 104 through various communication methods.
  • the user equipment 102 communicates with the terminal under test 104 through a wireless network or Bluetooth.
  • the embodiment of the present disclosure does not limit the communication mode.
  • the above-mentioned user equipment 102 may include but not limited to various smart phones, earphones, smart bracelets, smart watches and smart glasses;
  • the terminal 104 to be tested may include but not limited to various smart phones, tablet computers and earphones, smart watches, Portable wearable devices such as smart glasses.
  • a positioning method is provided, and the method is applied to the terminal under test in Figure 1 as an example for illustration, including the following steps:
  • Step 201 acquire inertial measurement data collected by at least one user equipment.
  • an IMU can be installed in the user equipment.
  • the IMU is a device for measuring the angular velocity and acceleration of an object.
  • An IMU generally includes an accelerometer and a gyroscope, which are used to measure the acceleration and angular velocity of an object in a three-dimensional space.
  • the user equipment collects inertial measurements through the IMU. data.
  • the terminal to be tested communicates with at least one user equipment, and acquires inertial measurement data from the user equipment.
  • the user wears the terminal to be tested and the user equipment at the same time, wherein the terminal to be tested is a smart phone, and the user equipment is a Bluetooth wireless headset.
  • the Bluetooth wireless headset is provided with an IMU, and the IMU collects inertial measurement data.
  • the smart phone communicates with the Bluetooth wireless headset through the Bluetooth protocol, and obtains the inertial measurement data from the Bluetooth wireless headset.
  • the at least one user equipment includes at least one of a smart phone, a headset, a smart bracelet, a smart watch, and smart glasses.
  • the inertial measurement data may also be acquired in other manners, and the embodiments of the present disclosure do not limit the acquisition manner.
  • Step 202 input the inertial measurement data into the speed prediction network to perform noise reduction and speed prediction, and obtain the speed of the terminal to be tested.
  • a pre-trained speed prediction network is set in the terminal to be tested, and the speed prediction network is trained by using sample inertial measurement data collected by multiple user equipments.
  • the terminal to be tested After the terminal to be tested obtains the inertial measurement data, it inputs the inertial measurement data into the speed prediction network, and the speed prediction network performs noise reduction on the inertial measurement data, and performs speed prediction based on the noise-reduced inertial measurement data, and the speed prediction network outputs The speed of the terminal under test.
  • the acceleration in the inertial measurement data at multiple collection moments is usually composed of an acceleration sequence, and then the acceleration sequence is input into the speed prediction network to obtain the speed of the terminal to be tested.
  • the number of accelerations in the acceleration sequence can be determined according to the frequency of data collection. For example, when the acquisition frequency is 200 Hz, the number of accelerations in the acceleration sequence may be 200. The embodiment of the present disclosure does not limit the number of accelerations.
  • Step 203 determine the positioning data of the terminal to be tested according to the speed of the terminal to be tested.
  • the positioning data includes at least one of position information and track information.
  • the terminal to be tested After the terminal to be tested acquires the velocity, it can perform integral processing according to the angular velocity in the velocity and inertial measurement data to obtain the positioning data of the terminal to be tested.
  • the above positioning method acquire the inertial measurement data collected by at least one user equipment, input the inertial measurement data into the speed prediction network for noise reduction and speed prediction, and obtain the speed of the terminal to be tested; determine the terminal to be tested according to the speed of the terminal to be tested location data.
  • the speed of the terminal to be tested is not obtained by the integral method, but the speed of the terminal to be tested is obtained by using the speed prediction network.
  • the speed prediction network is used Noise reduction can be achieved, thereby improving the accuracy of the speed.
  • the position information or trajectory information determined according to the speed will be more accurate, that is, the positioning accuracy is improved, and it can meet the needs of indoor positioning and navigation applications.
  • the above-mentioned process of acquiring the inertial measurement data collected by at least one user equipment may include the following steps:
  • Step 301 acquire raw inertial measurement data collected by at least one user equipment.
  • the user equipment collects the original inertial measurement data, and the terminal to be tested communicates with the user equipment to obtain the original inertial measurement data from the user equipment.
  • the Bluetooth wireless headset shakes with the shaking of the user's head, and the smart bracelet swings with the swing of the user's arm; Put it in your clothes pocket. Therefore, the original inertial measurement data obtained by the terminal under test from the user equipment is collected by the user equipment in different attitudes, and such data is not conducive to the speed prediction of the speed prediction network. Therefore, the terminal to be tested needs to perform data preprocessing on the original inertial measurement data, such as performing attitude conversion and coordinate conversion on the original inertial measurement data to obtain normalized inertial measurement data.
  • the above-mentioned process of performing attitude conversion and coordinate conversion on the original inertial measurement data to obtain the inertial measurement data may include:
  • Step 3021 converting the original inertial measurement data into first inertial measurement data of the user equipment in a standard attitude.
  • the attitude of the original inertial measurement data collected by the user equipment may be different from the standard attitude. Therefore, the original inertial measurement data is converted from the attitude at the time of data collection to the standard attitude to obtain the first inertial measurement data.
  • the attitude conversion process may include: obtaining the attitude angle of the user equipment according to the original inertial measurement data; according to the attitude angle of the user equipment, mapping the original inertial measurement data to the coordinate system of the user equipment under the standard attitude to obtain the first inertial measurement data .
  • the angular velocity in the original inertial measurement data to obtain the attitude angle of the user equipment; then determine the mapping relationship according to the attitude angle of the user equipment, and then map the original inertial measurement data to the user equipment under the standard attitude according to the mapping relationship In the coordinate system, the first inertial measurement data is obtained.
  • Step 3022 transforming the first inertial measurement data into the world coordinate system to obtain inertial measurement data.
  • the first inertial measurement data After obtaining the first inertial measurement data of the user equipment in the standard attitude, the first inertial measurement data can be converted to the world coordinate system according to the correspondence between the user equipment and the world coordinate system, and the speed prediction network can be obtained for speed prediction. Inertial measurement data required.
  • the process of transforming the coordinate system may further include: converting the first inertial measurement data to the terminal under test according to the correspondence between the coordinate system of the user equipment in the standard posture and the coordinate system of the terminal under test In the coordinate system, the second inertial measurement data is obtained; according to the corresponding relationship between the coordinate system of the terminal to be tested and the world coordinate system, the second inertial measurement data is converted to the world coordinate system to obtain the inertial measurement data.
  • the Bluetooth wireless headset is attached to the smart phone in advance, so that the smart phone can obtain the standard posture of the Bluetooth wireless headset, and obtain the correspondence between the coordinate system of the Bluetooth wireless headset in the standard posture and the coordinate system of the smart phone .
  • the user wears the Bluetooth wireless headset and the smartphone and moves, and the Bluetooth wireless headset transmits the collected raw inertial measurement data to the smartphone.
  • the smart phone first performs attitude conversion on the original inertial measurement data to obtain the first inertial measurement data; then according to the corresponding relationship between the coordinate system of the Bluetooth wireless headset under the standard attitude obtained in the above process and the coordinate system of the smart phone, the second The first inertial measurement data is transformed into the coordinate system of the smart phone to obtain the second inertial measurement data.
  • the smart phone converts the second inertial measurement data into the world coordinate system according to the corresponding relationship between the smart phone coordinate system and the world coordinate system, and obtains the inertial measurement data required by the speed prediction network for speed prediction.
  • the original inertial measurement data is converted into the first inertial measurement data of the user equipment in a standard attitude; the first inertial measurement data is converted into the world coordinate system to obtain the inertial measurement data.
  • the first inertial measurement data is obtained by performing attitude conversion on the original inertial measurement data, and then coordinate conversion is performed on the first inertial measurement data. Since the inertial measurement data is unified into the world coordinate system, it can be It is convenient for the speed prediction network to perform speed prediction, thereby improving the robustness of the speed prediction network.
  • the inertial measurement data collected by at least one user equipment includes inertial measurement data collected by multiple user equipments, as shown in FIG. 5 , this embodiment of the present disclosure may further include:
  • Step 204 Time-align the inertial measurement data collected by each user equipment according to the time stamps of the inertial measurement data collected by each user equipment, to obtain aligned inertial measurement data.
  • a Bluetooth wireless headset and a smart wristband collect inertial measurement data simultaneously. Therefore, after acquiring the inertial measurement data collected by multiple user equipments, the terminal under test performs time alignment on the inertial measurement data collected by multiple user equipments according to the time stamp, to obtain aligned inertial measurement data.
  • the smart phone time-aligns the inertial measurement data collected by the Bluetooth wireless headset with the inertial measurement data collected by the smart bracelet according to the time stamp, and obtains the aligned inertial measurement data.
  • multiple user equipments collect data at the same time, which can increase the amount of data, thereby improving the prediction speed and prediction efficiency of the speed prediction network.
  • the inertial measurement data after the alignment of the Bluetooth wireless headset and the alignment of the smart bracelet are input into the speed prediction network for noise reduction and speed prediction, and the speed of the smartphone is obtained.
  • time alignment is performed on the inertial measurement data collected by each user equipment according to the time stamps of the inertial measurement data collected by each user equipment, and the aligned Inertial measurement data; then input the aligned inertial measurement data into the speed prediction network for noise reduction and speed prediction to obtain the speed of the terminal to be tested.
  • data collection by multiple user equipment can expand the amount of data, thereby improving the prediction speed and prediction efficiency of the speed prediction network, but when using the inertial measurement data collected by multiple user equipment, time alignment is required Processing, so that the speed prediction network can perform speed prediction more accurately.
  • Step 401 acquiring track information of the terminal to be tested according to the speed of the terminal to be tested.
  • an integral operation can be performed according to the speed of the terminal to be tested and the angular velocity in the inertial measurement data to obtain the trajectory information of the terminal to be tested.
  • Step 402 matching the trajectory information with the preset scene map to obtain and output the location information of the terminal to be tested.
  • the terminal to be tested can obtain the preset scene map, and match the trajectory information with the preset scene map to obtain the location information of the terminal to be tested.
  • a smartphone acquires an indoor map of a museum, and by matching the trajectory information with the indoor map, the location information of the smartphone can be determined, that is, the specific location of the user in the museum can be determined.
  • the service information includes at least one of route guidance information and item introduction information.
  • the terminal under test can also obtain service information such as path guidance information and item introduction information according to the location information and the preset scene map, and output the service information in various ways.
  • service information such as path guidance information and item introduction information according to the location information and the preset scene map
  • the smart phone can obtain the route guidance information based on the location information and the indoor map of the museum, and then control the smart phone to display the route guidance information or control the Bluetooth wireless headset to play the route guidance information to guide the visitor to the next visiting point.
  • the smart phone can also obtain the item introduction information based on the location information and the indoor map of the museum, control the Bluetooth wireless headset to play the item introduction information, and introduce the cultural relics at the current location to the visitors.
  • the service information can also be combined with spatial audio, voice assistant and other functions to provide indoor voice navigation services for the blind.
  • the embodiment of the present disclosure does not limit the service information.
  • the trajectory information of the terminal to be tested is obtained according to the speed of the terminal to be tested, and the trajectory information is matched with the preset scene map to obtain and output the location information of the terminal to be tested; according to the location information of the terminal to be tested and the preset Scene map, get and output service information.
  • the terminal under test can provide users with services such as indoor voice navigation and voice navigation in large buildings according to the preset scene map and trajectory information, which fits with various typical application scenarios and improves the user experience .
  • the training process of the speed prediction network may include the following steps:
  • Step 501 acquire a training sample set.
  • the training sample set includes a plurality of sample inertial measurement data and labels corresponding to each sample inertial measurement data, labeled as actual speed.
  • Step 502 training the neural network model based on the training sample set to obtain an initial prediction network.
  • Input a sample inertial measurement data in the training sample set into the neural network model for training, and obtain the training result output by the neural network model; use the preset loss function to calculate the loss value between the training result and the label corresponding to the sample inertial measurement data ; If the loss value does not meet the preset convergence conditions, adjust the adjustable parameters in the neural network model, and input another sample inertial measurement data into the parameter-modified neural network model to continue training. Until the loss value between the training result output by the neural network model and the corresponding label meets the preset convergence condition, the training is ended, and the neural network model at the end of the training is determined as the initial prediction network.
  • Step 503 Test the initial prediction network based on the training sample set, and determine multiple test trajectories according to the prediction speed obtained from the test.
  • the initial prediction network is tested using the sample inertial measurement data in the training sample set that did not participate in the training. Specifically, multiple sample inertial measurement data that have not participated in training are sequentially input into the initial prediction network, and the predicted speed corresponding to each sample inertial measurement data output by the initial prediction network is obtained. Integral calculations are performed according to multiple predicted speeds to obtain multiple test tracks.
  • Step 504 if the error among the multiple test trajectories is smaller than the preset error value, determine the initial prediction network as the speed prediction network.
  • the training sample set is obtained; the training of the neural network model is carried out based on the training sample set to obtain an initial prediction network; the initial prediction network is tested based on the training sample set, and a plurality of test tracks are determined according to the prediction speed obtained by the test; When the errors among the multiple test trajectories are smaller than the preset error value, the initial prediction network is determined as the speed prediction network.
  • the training sample set is used to train the neural network model and test the initial prediction network, and the training ends when the prediction accuracy of the initial prediction network is high enough to ensure that the speed prediction network obtained by training is accurate. sex.
  • the above-mentioned process of obtaining a training sample set may include the following steps:
  • Step 5011 respectively acquire a plurality of original inertial measurement data collected by the data collection device and the actual trajectory collected by the annotation collection device.
  • a plurality of original inertial measurement data collected by the data collection device is obtained, and the actual trajectory collected by the labeling device is obtained.
  • obtain multiple raw inertial measurement data collected by data collection devices such as Bluetooth wireless headsets and smart bracelets
  • obtain data collected by annotation collection devices such as smartphones or head-mounted display devices (AR glasses) deployed with visual odometry (VIO) algorithms. actual trajectory.
  • data collection devices such as Bluetooth wireless headsets and smart bracelets
  • annotation collection devices such as smartphones or head-mounted display devices (AR glasses) deployed with visual odometry (VIO) algorithms.
  • VIO visual odometry
  • the data collection equipment and label collection equipment can be placed together in the calibration device, and the calibration device can be used to calibrate the data collection time of the data collection equipment and the label collection equipment, and according to the difference between the data collection equipment and the label collection equipment The relative position between them establishes the coordinate mapping relationship between them.
  • the data collection device and the labeling collection device need to be worn together, and the labeling collection device needs to be close to the collector, so as to ensure that the 6Dof (6 degrees of freedom output by the labeling collection device, three degrees of freedom describe the position information, Three degrees of freedom describe attitude information)
  • the trajectory is exactly the same as that of the human body.
  • Data collection can adopt multiple people, multi-device types, multi-batch, and multi-scenario collection methods to ensure the randomness, richness, and data volume of the original inertial measurement data.
  • the embodiment of the present disclosure does not limit the data collection device and the annotation collection device, which can be set according to actual conditions.
  • Step 5012 performing attitude transformation and coordinate transformation on each original inertial measurement data respectively, to obtain sample inertial measurement data corresponding to each original inertial measurement data.
  • attitude conversion on each original inertial measurement data so as to convert each original inertial measurement data from the coordinate system of data acquisition attitude to the coordinate system of standard attitude; and then perform coordinate conversion on each converted inertial measurement data
  • the converted inertial measurement data are converted from the standard attitude coordinate system to the world coordinate system to obtain sample inertial measurement data corresponding to each original inertial measurement data.
  • the inertial measurement data under the standard attitude can be converted into the coordinate system of the label collection equipment, and then the coordinates of the label collection equipment can be used to system to the world coordinate system.
  • attitude conversion and coordinate conversion process reference may be made to the above-mentioned embodiments, and the embodiments of the present disclosure will not be repeated here.
  • Step 5013 After time aligning the collection time of multiple sample inertial measurement data with the collection time of the actual trajectory, determine the actual velocity corresponding to each sample inertial measurement data.
  • the time alignment can be performed according to the acquisition time of the sample inertial measurement data and the acquisition time of the actual trajectory, so that , the actual velocity corresponding to the inertial measurement data of each sample can be obtained.
  • Step 5014 obtain a training sample set according to the multiple sample inertial measurement data and the actual velocity corresponding to each sample inertial measurement data.
  • a training sample set is formed from the plurality of sample inertial measurement data and the actual speed corresponding to each sample inertial measurement data.
  • the multiple original inertial measurement data collected by the data acquisition device and the actual trajectory collected by the annotation acquisition device are respectively obtained; attitude conversion and coordinate conversion are performed on each original inertial measurement data respectively, and samples corresponding to each original inertial measurement data are obtained Inertial measurement data; after time aligning the acquisition time of multiple sample inertial measurement data with the acquisition time of the actual trajectory, determine the actual speed corresponding to each sample inertial measurement data; according to multiple sample inertial measurement data and each sample inertial measurement data The corresponding actual speed obtains the training sample set.
  • the coordinate mapping relationship between the data collection equipment and the annotation collection equipment is established, and the time calibration of the two is performed, so that after the data collection, the inertia can be adjusted according to the coordinate mapping relationship.
  • the measurement data is normalized, and the corresponding relationship between the sample inertial measurement data and the actual speed can also be obtained, so as to obtain the training sample set, which is a good preparation for the training speed prediction network.
  • steps in the flow charts of FIG. 2 to FIG. 9 are shown sequentially as indicated by the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in FIGS. 2 to 9 may include multiple steps or stages, and these steps or stages are not necessarily performed at the same time, but may be performed at different times. The steps or stages The order of execution is not necessarily performed sequentially, but may be performed alternately or alternately with other steps or at least a part of steps or stages in other steps.
  • a positioning device comprising:
  • the speed prediction module 602 is used to input the inertial measurement data into the speed prediction network to perform noise reduction and speed prediction to obtain the speed of the terminal to be tested; the speed prediction network is obtained by training the sample inertial measurement data collected by multiple user equipments;
  • the positioning data determining module 603 is configured to determine the positioning data of the terminal to be tested according to the speed of the terminal to be tested.
  • the above-mentioned inertial data acquisition module 601 includes:
  • the above conversion sub-module is specifically used to obtain the attitude angle of the user equipment according to the original inertial measurement data; according to the attitude angle of the user equipment, map the original inertial measurement data to the coordinate system of the user equipment under the standard attitude , get the first inertial measurement data.
  • the above conversion sub-module is specifically used to convert the first inertial measurement data to the coordinate system of the terminal under test according to the corresponding relationship between the coordinate system of the user equipment under the standard attitude and the coordinate system of the terminal under test.
  • the second inertial measurement data is obtained; according to the corresponding relationship between the coordinate system of the terminal to be tested and the world coordinate system, the second inertial measurement data is converted into the world coordinate system to obtain the inertial measurement data.
  • the inertial measurement data collected by at least one user equipment includes inertial measurement data collected by multiple user equipments, as shown in FIG. 11 , the apparatus further includes:
  • the alignment module 604 is configured to perform time alignment on the inertial measurement data collected by each user equipment according to the time stamps of the inertial measurement data collected by each user equipment, to obtain aligned inertial measurement data;
  • the above-mentioned speed prediction module 603 is specifically used to input the aligned inertial measurement data into the speed prediction network for noise reduction and speed prediction to obtain the speed of the terminal to be tested.
  • the above positioning data determination module 603 is used to obtain the trajectory information of the terminal to be tested according to the speed of the terminal to be tested; match the trajectory information with the preset scene map to obtain and output the location information of the terminal to be tested .
  • the device further includes:
  • the service information output module 605 is configured to obtain and output service information according to the location information of the terminal to be tested and the preset scene map; wherein, the service information includes at least one of path guidance information and item introduction information.
  • the device further includes:
  • the sample acquisition module 606 is used to obtain a training sample set; wherein, the training sample set includes a plurality of sample inertial measurement data and labels corresponding to each sample inertial measurement data, marked as actual speed;
  • the training module 607 is used to train the neural network model based on the training sample set to obtain the initial prediction network;
  • the testing module 608 is used to test the initial prediction network based on the training sample set, and determine a plurality of test tracks according to the prediction speed obtained by the test;
  • the network determination module 609 is configured to determine the initial prediction network as the speed prediction network when the error between the multiple test trajectories is less than a preset error value.
  • the above-mentioned sample acquisition module 606 is specifically used to respectively acquire a plurality of original inertial measurement data collected by the data acquisition device and the actual trajectory collected by the annotation acquisition device; respectively perform attitude conversion and coordinates on each original inertial measurement data Conversion to obtain the sample inertial measurement data corresponding to each original inertial measurement data; after time aligning the collection time of multiple sample inertial measurement data with the collection time of the actual trajectory, determine the actual speed corresponding to each sample inertial measurement data; according to multiple Each sample inertial measurement data and the actual velocity corresponding to each sample inertial measurement data are used to obtain a training sample set.
  • At least one user equipment includes at least one of a smart phone, a headset, a smart bracelet, a smart watch, and smart glasses.
  • Each module in the above positioning device can be fully or partially realized by software, hardware and a combination thereof.
  • the above-mentioned modules may be embedded in or independent of the processor in the terminal in the form of hardware, and may also be stored in the memory of the terminal in the form of software, so that the processor can invoke and execute the corresponding operations of the above-mentioned modules.
  • a terminal is provided, and its internal structure diagram may be as shown in FIG. 14 .
  • the terminal includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Wherein, the processor of the terminal is used to provide calculation and control capabilities.
  • the memory of the terminal includes a non-volatile storage medium and an internal memory.
  • the non-volatile storage medium stores an operating system and computer programs.
  • the internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium.
  • the communication interface of the terminal is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, an operator network, NFC (Near Field Communication) or other technologies.
  • the display screen of the terminal may be a liquid crystal display screen or an electronic ink display screen
  • the input device of the terminal may be a touch layer covered on the display screen, or a button, a trackball or a touch pad provided on the terminal shell, or It is an external keyboard, trackpad or mouse.
  • a terminal including a memory and a processor, a computer program is stored in the memory, and the processor implements the following steps when executing the computer program:
  • the speed prediction network is obtained by training the sample inertial measurement data collected by multiple user devices;
  • Attitude transformation and coordinate transformation are performed on the original inertial measurement data to obtain the inertial measurement data.
  • the original inertial measurement data is mapped to the coordinate system of the user equipment under the standard attitude to obtain the first inertial measurement data.
  • the second inertial measurement data is converted to the world coordinate system to obtain the inertial measurement data.
  • the inertial measurement data collected by at least one user equipment includes inertial measurement data collected by multiple user equipments
  • the processor further implements the following steps when executing the computer program:
  • service information is obtained and output; wherein, the service information includes at least one of path guidance information and item introduction information.
  • the training sample set includes a plurality of sample inertial measurement data and labels corresponding to each sample inertial measurement data, labeled as actual speed;
  • the initial prediction network is determined as the speed prediction network.
  • a training sample set is obtained according to the plurality of sample inertial measurement data and the actual velocity corresponding to each sample inertial measurement data.
  • the at least one user equipment includes at least one of a smart phone, a headset, a smart bracelet, a smart watch, and smart glasses.
  • a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
  • the positioning data of the terminal to be tested is determined according to the speed of the terminal to be tested.
  • Attitude transformation and coordinate transformation are performed on the original inertial measurement data to obtain the inertial measurement data.
  • the original inertial measurement data is mapped to the coordinate system of the user equipment under the standard attitude to obtain the first inertial measurement data.
  • the second inertial measurement data is converted to the world coordinate system to obtain the inertial measurement data.
  • the inertial measurement data collected by at least one user equipment includes inertial measurement data collected by multiple user equipments
  • the computer program further implements the following steps when executed by the processor:
  • the training sample set includes a plurality of sample inertial measurement data and labels corresponding to each sample inertial measurement data, labeled as actual speed;
  • the initial prediction network is determined as the speed prediction network.
  • a training sample set is obtained according to the plurality of sample inertial measurement data and the actual velocity corresponding to each sample inertial measurement data.
  • the at least one user equipment includes at least one of a smart phone, a headset, a smart bracelet, a smart watch, and smart glasses.
  • Non-volatile memory may include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory or optical memory, etc.
  • Volatile memory can include Random Access Memory (RAM) or external cache memory.
  • RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Navigation (AREA)

Abstract

本申请涉及一种定位方法、装置、终端和存储介质。所述方法包括：获取至少一个用户设备采集的惯性测量数据；将所述惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；所述速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；根据所述待测终端的速度确定所述待测终端的定位数据。采用本方法能够提高定位精度，满足室内定位与导航应用需求。

Description

定位方法、装置、终端和存储介质
相关申请的交叉引用
本申请要求2021年10月26日申请的，申请号为2021112487812，名称为“定位方法、装置、终端和存储介质”的中国专利申请的优先权，在此将其全文引入作为参考。
技术领域
本申请涉及定位技术领域，特别是涉及一种定位方法、装置、终端和存储介质。
背景技术
随着科技的飞速发展，人们对智能终端的功能需求越来越多，其中利用智能终端进行定位和导航就是需求之一。
传统技术中，通过智能终端中安装的惯性测量单元(Inertial Measurement Unit，IMU)采集惯性导航数据，根据惯性导航数据确定智能终端的位置信息或运动轨迹。然而，上述定位方法的定位精度难以满足室内定位与导航应用的需求。
发明内容
基于此，有必要针对上述技术问题，提供一种能够提高定位精度，满足室内定位与导航应用需求的定位方法、装置、终端和存储介质。
第一方面，本公开实施例提供了一种定位方法，该方法包括：
获取至少一个用户设备采集的惯性测量数据；
将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
根据待测终端的速度确定待测终端的定位数据。
在其中一个实施例中，上述获取至少一个用户设备采集的惯性测量数据，包括：
获取至少一个用户设备采集的原始惯性测量数据；
对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据。
在其中一个实施例中，上述对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据，包括：
将原始惯性测量数据转换为用户设备在标准姿态下的第一惯性测量数据；
将第一惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在其中一个实施例中，上述将原始惯性测量数据转换为用户设备在标准姿态下的第一惯性测量数据，包括：
根据原始惯性测量数据获取用户设备的姿态角；
根据用户设备的姿态角，将原始惯性测量数据映射至用户设备在标准姿态下的坐标系中，得到第一惯性测量数据。
在其中一个实施例中，上述将第一惯性测量数据转换至世界坐标系下，得到惯性测量数据，包括：
根据用户设备在标准姿态下的坐标系与待测终端的坐标系之间的对应关系，将第一惯性测量数据转换至待测终端的坐标系中，得到第二惯性测量数据；
根据待测终端的坐标系与世界坐标系之间的对应关系，将第二惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在其中一个实施例中，至少一个用户设备采集的惯性测量数据包括多个用户设备采集的惯性测量数据，该方法还包括：
根据各用户设备采集惯性测量数据的时间戳，对各用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据；
将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度，包括：
将对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。
在其中一个实施例中，上述根据待测终端的速度确定待测终端的定位数据，包括：
根据待测终端的速度获取待测终端的轨迹信息；
将轨迹信息与预设场景地图进行匹配，得到并输出待测终端的位置信息。
在其中一个实施例中，该方法还包括：
根据待测终端的位置信息和预设场景地图，得到并输出服务信息；其中，服务信息包括路径导引信息和物品介绍信息中的至少一种。
在其中一个实施例中，上述速度预测网络的训练过程包括：
获取训练样本集；其中，训练样本集包括多个样本惯性测量数据和各样本惯性测量数据对应的标注，标注为实际速度；
基于训练样本集进行神经网络模型的训练，得到初始预测网络；
基于训练样本集对初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹；
在多个测试轨迹之间的误差小于预设误差值的情况下，将初始预测网络确定为速度预测网络。
在其中一个实施例中，上述获取训练样本集，包括：
分别获取数据采集设备采集的多个原始惯性测量数据和标注采集设备采集的实际轨迹；
分别对各原始惯性测量数据进行姿态转换和坐标转换，得到各原始惯性测量数据对应的样本惯性测量数据；
将多个样本惯性测量数据的采集时刻与实际轨迹的采集时刻进行时间对准后，确定各样本惯性测量数据对应的实际速度；
根据多个样本惯性测量数据和各样本惯性测量数据对应的实际速度得到训练样本集。
在其中一个实施例中，至少一个用户设备包括智能手机、耳机、智能手环、智能手表 和智能眼镜中的至少一种。
第二方面，本公开实施例提供了一种定位装置，该装置包括：
惯性数据获取模块，用于获取至少一个用户设备采集的惯性测量数据；
速度预测模块，用于将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
定位数据确定模块，用于根据待测终端的速度确定待测终端的定位数据。
在其中一个实施例中，上述惯性数据获取模块，包括：
原始数据获取子模块，用于获取至少一个用户设备采集的原始惯性测量数据；
转换子模块，用于对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据。
在其中一个实施例中，上述转换子模块，具体用于将原始惯性测量数据转换为用户设备在标准姿态下的第一惯性测量数据；将第一惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在其中一个实施例中，上述转换子模块，具体用于根据原始惯性测量数据获取用户设备的姿态角；根据用户设备的姿态角，将原始惯性测量数据映射至用户设备在标准姿态下的坐标系中，得到第一惯性测量数据。
在其中一个实施例中，上述转换子模块，具体用于根据用户设备在标准姿态下的坐标系与待测终端的坐标系之间的对应关系，将第一惯性测量数据转换至待测终端的坐标系中，得到第二惯性测量数据；根据待测终端的坐标系与世界坐标系之间的对应关系，将第二惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在其中一个实施例中，至少一个用户设备采集的惯性测量数据包括多个用户设备采集的惯性测量数据，该装置还包括：
对准模块，用于根据各用户设备采集惯性测量数据的时间戳，对各用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据；
上述速度预测模块，具体用于将对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。
在其中一个实施例中，上述定位数据确定模块，用于根据待测终端的速度获取待测终端的轨迹信息；将轨迹信息与预设场景地图进行匹配，得到并输出待测终端的位置信息。
在其中一个实施例中，该装置还包括：
服务信息输出模块，用于根据待测终端的位置信息和预设场景地图，得到并输出服务信息；其中，服务信息包括路径导引信息和物品介绍信息中的至少一种。
在其中一个实施例中，该装置还包括：
样本获取模块，用于获取训练样本集；其中，训练样本集包括多个样本惯性测量数据和各样本惯性测量数据对应的标注，标注为实际速度；
训练模块，用于基于训练样本集进行神经网络模型的训练，得到初始预测网络；
测试模块，用于基于训练样本集对初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹；
网络确定模块，用于在多个测试轨迹之间的误差小于预设误差值的情况下，将初始预测网络确定为速度预测网络。
在其中一个实施例中，上述样本获取模块，具体用于分别获取数据采集设备采集的多个原始惯性测量数据和标注采集设备采集的实际轨迹；分别对各原始惯性测量数据进行姿态转换和坐标转换，得到各原始惯性测量数据对应的样本惯性测量数据；将多个样本惯性测量数据的采集时刻与实际轨迹的采集时刻进行时间对准后，确定各样本惯性测量数据对应的实际速度；根据多个样本惯性测量数据和各样本惯性测量数据对应的实际速度得到训练样本集。
在其中一个实施例中，至少一个用户设备包括智能手机、耳机、智能手环、智能手表和智能眼镜中的至少一种。
第三方面，本公开实施例提供了一种终端，包括存储器和处理器，所述存储器存储有计算机程序，所述处理器执行所述计算机程序时实现以下步骤：
获取至少一个用户设备采集的惯性测量数据；
将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
根据待测终端的速度确定待测终端的定位数据。
第四方面，本公开实施例提供了一种计算机可读存储介质，其上存储有计算机程序，所述计算机程序被处理器执行时实现以下步骤：
获取至少一个用户设备采集的惯性测量数据；
将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
根据待测终端的速度确定待测终端的定位数据。
上述定位方法、装置、终端和存储介质，获取至少一个用户设备采集的惯性测量数据，将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；根据待测终端的速度确定待测终端的定位数据。通过本公开实施例，在对待测终端进行定位的过程中，并未采取积分方式获得待测终端的速度，而是采用速度预测网络获得待测终端的速度，采用速度预测网络与采用积分相比可以实现降噪，从而提高速度的准确性，这样，根据速度确定的位置信息或轨迹信息就会更加准确，即提高了定位精度，可以满足室内定位与导航应用的需求。
附图说明
图1为一个实施例中定位方法的应用环境图；
图2为一个实施例中定位方法的流程示意图；
图3为一个实施例中获取至少一个用户设备采集的惯性测量数据步骤的流程示意图；
图4为一个实施例中对原始惯性测量数据进行姿态转换和坐标转换步骤的流程示意图；
图5为另一个实施例中定位方法的流程示意图；
图6为一个实施例中根据待测终端的速度确定待测终端的定位数据步骤的流程示意图；
图7为一个实施例中输出服务信息步骤的流程示意图；
图8为一个实施例中速度预测网络的训练步骤的流程示意图；
图9为一个实施例中获取训练样本集步骤的流程示意图；
图10为一个实施例中定位装置的结构框图之一；
图11为一个实施例中定位装置的结构框图之二；
图12为一个实施例中定位装置的结构框图之三；
图13为一个实施例中定位装置的结构框图之四；
图14为一个实施例中终端的内部结构图。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处描述的具体实施例仅仅用以解释本申请，并不用于限定本申请。
首先，在具体介绍本公开实施例的技术方案之前，先对本公开实施例基于的技术背景或者技术演进脉络进行介绍。通常情况下，通过智能终端中安装的惯性测量单元(Inertial Measurement Unit，IMU)采集惯性导航数据，根据惯性导航数据确定智能终端的位置信息或运动轨迹。其中，根据惯性导航数据确定智能终端的位置信息或运动轨迹常采用双积分法，这种方式容易产生轨迹误差，影响定位精度，使得定位精度难以满足室内定位与导航应用的需求。
而本申请所提供的技术方案中，先获取至少一个用户设备采集的惯性测量数据；然后将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；最后根据待测终端的速度确定待测终端的定位数据。通过本申请，不采用双积分法而是采用速度预测网络得到待测终端的速度，进而得到待测终端的定位数据，可以减少轨迹误差、提高定位精度，满足室内定位与导航应用的需求。
本申请提供的定位方法，可以应用于如图1所示的应用环境中。该应用环境包括用户设备102和待测终端104。其中，用户设备102可以通过多种通信方式与待测终端104进行通信。例如，用户设备102通过无线网络，或者蓝牙与待测终端104进行通信。本公开实施例对通信方式不做限定。上述用户设备102可以包括但不限于是各种智能手机、耳机、智能手环、智能手表和智能眼镜；待测终端104可以包括但不限于是各种智能手机、平板电脑和耳机、智能手表、智能眼镜等便携式可穿戴设备。
在一个实施例中，如图2所示，提供了一种定位方法，以该方法应用于图1中的待测 终端为例进行说明，包括以下步骤：
步骤201，获取至少一个用户设备采集的惯性测量数据。
其中，用户设备内可以设置IMU，IMU是测量物体角速度和加速度的装置，一个IMU一般包含加速度计和陀螺仪，分别用于测量物体在三维空间中的加速度和角速度，用户设备通过IMU采集惯性测量数据。
待测终端与至少一个用户设备进行通信，从用户设备获取到惯性测量数据。例如，用户同时佩戴待测终端和用户设备，其中，待测终端为智能手机、用户设备为蓝牙无线耳机。蓝牙无线耳机中设置有IMU，由IMU采集惯性测量数据。智能手机与蓝牙无线耳机通过蓝牙协议进行通信，从蓝牙无线耳机获取惯性测量数据。
可选地，至少一个用户设备包括智能手机、耳机、智能手环、智能手表和智能眼镜中的至少一种。在实际应用中，还可以通过其他方式获取惯性测量数据，本公开实施例对获取方式不做限定。
步骤202，将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。
其中，待测终端中设置有预先训练的速度预测网络，该速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到。
待测终端在获取到惯性测量数据后，将惯性测量数据输入到速度预测网络中，速度预测网络对惯性测量数据进行降噪，并依据降噪后的惯性测量数据进行速度预测，速度预测网络输出待测终端的速度。
在实际应用中，通常会将多个采集时刻的惯性测量数据中的加速度组成加速度序列，然后将加速度序列输入到速度预测网络中，得到待测终端的速度。其中，加速度序列中的加速度数量可以根据数据采集频率确定。例如，在采集频率为200Hz的情况下，加速度序列中的加速度数量可以为200个。本公开实施例对加速度数量不做限定。
需要说明的是，上述速度和加速度均为矢量，包括大小和方向。
步骤203，根据待测终端的速度确定待测终端的定位数据。
其中，定位数据包括位置信息和轨迹信息中的至少一种。
待测终端获取到速度后，可以根据速度和惯性测量数据中的角速度进行积分处理，得到待测终端的定位数据。
上述定位方法中，获取至少一个用户设备采集的惯性测量数据，将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；根据待测终端的速度确定待测终端的定位数据。通过本公开实施例，在对待测终端进行定位的过程中，并未采取积分方式获得待测终端的速度，而是采用速度预测网络获得待测终端的速度，采用速度预测网络与采用积分相比可以实现降噪，从而提高速度的准确性，这样，根据速度确定的位置信息或轨迹信息就会更加准确，即提高了定位精度，可以满足室内定位与导航应用的需求。
在一个实施例中，如图3所示，上述获取至少一个用户设备采集的惯性测量数据的过 程，可以包括如下步骤：
步骤301，获取至少一个用户设备采集的原始惯性测量数据。
用户设备采集到的是原始惯性测量数据，待测终端与用户设备进行通信，从用户设备获取原始惯性测量数据。
步骤302，对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据。
由于用户设备在不同时刻可能处于不同的姿态，比如，蓝牙无线耳机随用户头部的晃动而晃动，智能手环随用户手臂的摆动而摆动；又比如，智能手机处于用户手持状态，或者被用户放置在衣服口袋中。因此，待测终端从用户设备获取到的原始惯性测量数据是用户设备在不同姿态下采集到的，这样的数据不利于速度预测网络进行速度预测。所以，待测终端需要对原始惯性测量数据进行数据预处理，比如对原始惯性测量数据进行姿态转换和坐标转换，得到归一化的惯性测量数据。
上述实施例中，获取至少一个用户设备采集的原始惯性测量数据；对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据。通过本公开实施例，对惯性测量数据进行姿态转换和坐标转换，就是将任意时刻、任意设备朝向、任意行动状态下采集的原始惯性测量数据投影至统一的世界坐标系下，以便速度预测网络进行速度预测，从而提高速度预测网络的鲁棒性。
在一个实施例中，如图4所示，上述对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据的过程，可以包括：
步骤3021，将原始惯性测量数据转换为用户设备在标准姿态下的第一惯性测量数据。
在实际应用中，用户设备采集原始惯性测量数据的姿态可能与标准姿态不同，因此，先将原始惯性测量数据从数据采集时的姿态转换到标准姿态下，得到第一惯性测量数据。
姿态转换的过程可以包括：根据原始惯性测量数据获取用户设备的姿态角；根据用户设备的姿态角，将原始惯性测量数据映射至用户设备在标准姿态下的坐标系中，得到第一惯性测量数据。
例如，对原始惯性测量数据中的角速度进行积分得到用户设备的姿态角；然后根据用户设备的姿态角确定映射关系，接着，根据该映射关系将原始惯性测量数据映射到用户设备在标准姿态下的坐标系中，得到第一惯性测量数据。
步骤3022，将第一惯性测量数据转换至世界坐标系下，得到惯性测量数据。
得到用户设备在标准姿态下的第一惯性测量数据后，可以根据用户设备和世界坐标系之间的对应关系，将第一惯性测量数据转换到世界坐标系下，得到速度预测网络进行速度预测所需的惯性测量数据。
在其中一个实施例中，坐标系转换的过程还可以包括：根据用户设备在标准姿态下的坐标系与待测终端的坐标系之间的对应关系，将第一惯性测量数据转换至待测终端的坐标系中，得到第二惯性测量数据；根据待测终端的坐标系与世界坐标系之间的对应关系，将第二惯性测量数据转换至世界坐标系下，得到惯性测量数据。
例如，预先将蓝牙无线耳机与智能手机贴靠，使得智能手机可以获取到蓝牙无线耳机的标准姿态，并获取到蓝牙无线耳机在标准姿态下的坐标系与智能手机的坐标系之间的对应关系。接着，用户佩戴蓝牙无线耳机和智能手机并移动，蓝牙无线耳机将采集到的原始惯性测量数据传输到智能手机。智能手机先对原始惯性测量数据进行姿态转换，得到第一惯性测量数据；再根据上述过程中获得的蓝牙无线耳机在标准姿态下的坐标系与智能手机的坐标系之间的对应关系，将第一惯性测量数据转换到智能手机的坐标系下，得到第二惯性测量数据。之后，智能手机再根据智能手机的坐标系与世界坐标系之间的对应关系，将第二惯性测量数据转换到世界坐标系下，得到速度预测网络进行速度预测所需的惯性测量数据。
上述实施例中，将原始惯性测量数据转换为用户设备在标准姿态下的第一惯性测量数据；将第一惯性测量数据转换至世界坐标系下，得到惯性测量数据。通过本公开实施例，先对原始惯性测量数据进行姿态转换得到第一惯性测量数据，再对第一惯性测量数据进行坐标转换惯性测量数据，由于将惯性测量数据统一到了世界坐标系下，因此可以方便速度预测网络进行速度预测，从而提高速度预测网络的鲁棒性。
在一个实施例中，至少一个用户设备采集的惯性测量数据包括多个用户设备采集的惯性测量数据，如图5所示，本公开实施例还可以包括：
步骤204，根据各用户设备采集惯性测量数据的时间戳，对各用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据。
在实际应用中，可能采集惯性测量数据的用户设备有多个。例如，蓝牙无线耳机和智能手环同时采集惯性测量数据。因此，在待测终端在获取到多个用户设备采集的惯性测量数据后，根据时间戳对多个用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据。
例如，智能手机根据时间戳将蓝牙无线耳机采集到的惯性测量数据与智能手环采集到的惯性测量数据进行时间对准，得到对准后的惯性测量数据。
可以理解地，多个用户设备同时进行数据采集，可以增加数据数量，进而提高速度预测网络的预测速度和预测效率。
对应地，步骤203可以包括：将对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。
例如，将蓝牙无线耳机对准后的惯性测量数据和智能手环对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到智能手机的速度。
上述实施例中，在多个用户设备采集的惯性测量数据的情况下，根据各用户设备采集惯性测量数据的时间戳，对各用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据；再将对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。通过本公开实施例，由多个用户设备进行数据采集可以扩展数据量，从而提高速度预测网络的预测速度和预测效率，而使用多个用户设备采集到的惯性测量数 据时，需要进行时间对准处理，以便速度预测网络可以更为准确地进行速度预测。
在一个实施例中，如图6所示，上述根据待测终端的速度确定待测终端的定位数据的过程，可以包括如下步骤：
步骤401，根据待测终端的速度获取待测终端的轨迹信息。
在速度预测网络输出待测终端的速度后，可以根据待测终端的速度和惯性测量数据中的角速度进行积分运算，得到待测终端的轨迹信息。
步骤402，将轨迹信息与预设场景地图进行匹配，得到并输出待测终端的位置信息。
在室内定位和导航应用的场景下，待测终端可以获取到预设场景地图，并将轨迹信息与预设场景地图进行匹配，得到待测终端的位置信息。例如，智能手机获取到某博物馆的室内地图，将轨迹信息与该室内地图进行匹配，可以确定智能手机的位置信息，即确定用户在博物馆中的具体位置。
之后，待测终端可以通过多种方式输出待测终端的位置信息。例如，可以通过智能手机、智能眼镜等显示位置信息，也可以通过智能手机、蓝牙无线耳机等播放位置信息，还可以通过智能手环、智能手表振动来提示位置信息。本公开实施例对输出方式不做限定，可以根据实际情况进行设置。
在上述实施例的基础上，如图7所示，本公开实施例还可以包括：
步骤403，根据待测终端的位置信息和预设场景地图，得到并输出服务信息。
其中，服务信息包括路径导引信息和物品介绍信息中的至少一种。
待测终端还可以根据位置信息和预设场景地图，得到路径导引信息、物品介绍信息等服务信息，并采用多种方式输出服务信息。
例如，智能手机可以根据位置信息和博物馆的室内地图得到路径引导信息，之后，控制智能手机显示路径导引信息或控制蓝牙无线耳机播放路径导引信息，指引参观者前往下一个参观点。智能手机也可以根据位置信息和博物馆的室内地图得到物品介绍信息，控制蓝牙无线耳机播放物品介绍信息，为参观者介绍当前位置的文物。
在其中一个实施例中，服务信息还可以结合空间音频、语音助手等功能，为盲人提供室内语音导航服务。本公开实施例对服务信息不做限定。
上述实施例中，根据待测终端的速度获取待测终端的轨迹信息，将轨迹信息与预设场景地图进行匹配，得到并输出待测终端的位置信息；根据待测终端的位置信息和预设场景地图，得到并输出服务信息。通过本公开实施例，待测终端可以根据预设场景地图和轨迹信息为用户提供室内语音导航、大型建筑物内语音导览等服务，与多种典型的应用场景契合，提升了用户的使用体验。
在一个实施例中，如图8所示，速度预测网络的训练过程可以包括如下步骤：
步骤501，获取训练样本集。
其中，训练样本集包括多个样本惯性测量数据和各样本惯性测量数据对应的标注，标注为实际速度。
采用多个用户设备进行数据采集得到多个样本惯性测量数据和各样本惯性测量数据对应的实际速度，由多个样本惯性测量数据和各样本惯性测量数据对应的实际速度组成训练样本集。
步骤502，基于训练样本集进行神经网络模型的训练，得到初始预测网络。
将训练样本集中的一个样本惯性测量数据输入到神经网络模型中进行训练，得到神经网络模型输出的训练结果；利用预设损失函数计算训练结果与该样本惯性测量数据对应的标注之间的损失值；若该损失值不符合预设收敛条件，则调整神经网络模型中的可调参数，并将另一个样本惯性测量数据输入到参数修改后的神经网络模型中继续进行训练。直到神经网络模型输出的训练结果与对应的标注之间的损失值符合预设收敛条件时，结束训练，并将结束训练时的神经网络模型确定为初始预测网络。
步骤503，基于训练样本集对初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹。
采用训练样本集中未参与训练的样本惯性测量数据对初始预测网络进行测试。具体地，将多个未参与训练的样本惯性测量数据依次输入到初始预测网络中，得到初始预测网络输出的各样本惯性测量数据对应的预测速度。根据多个预测速度进行积分运算得到多个测试轨迹。
步骤504，在多个测试轨迹之间的误差小于预设误差值的情况下，将初始预测网络确定为速度预测网络。
计算多个测试轨迹之间的误差，如果多个测试轨迹之间的误差小于预设误差值，表明初始预测网络的预测准确性较高，则将初始预测网络确定为速度预测网络。如果多个测试轨迹之间的误差大于或等于预设误差值，表明初始预测网络的预测准确性不够高，则可以重复步骤502-504，直到多个测试轨迹之间的误差小于预设误差值时结束训练，得到速度预测网络。
例如，预设误差值为行走100m产生的误差在3m以内，如果多个测试轨迹之间的误差小于上述预测误差值，则将初始预测网络确定为速度预测网络。本公开实施例对预测误差值不做限定。
上述实施例中，获取训练样本集；基于训练样本集进行神经网络模型的训练，得到初始预测网络；基于训练样本集对初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹；在多个测试轨迹之间的误差小于预设误差值的情况下，将初始预测网络确定为速度预测网络。通过本公开实施例，利用训练样本集进行神经网络模型的训练和初始预测网络的测试，在初始预测网络的预测准确性足够高的情况下结束训练，以保证训练得到的速度预测网络的预测准确性。
在一个实施例中，如图9所示，上述获取训练样本集的过程，可以包括如下步骤：
步骤5011，分别获取数据采集设备采集的多个原始惯性测量数据和标注采集设备采集的实际轨迹。
在获取训练样本集的过程中，获取数据采集设备采集的多个原始惯性测量数据，获取标注设备采集实际轨迹。
例如，获取蓝牙无线耳机、智能手环等数据采集设备采集的多个原始惯性测量数据，获取部署有视觉里程计(VIO)算法的智能手机或头戴显示设备(AR眼镜)等标注采集设备采集的实际轨迹。
在数据采集前，可以将数据采集设备与标注采集设备一同放置在校准装置中，利用校准装置对数据采集设备和标注采集设备进行数据采集时间的校准处理，并根据数据采集设备和标注采集设备之间的相对位置，建立两者之间的坐标映射关系。
在数据采集过程中，数据采集设备与标注采集设备需一同佩戴，并且，标注采集设备需与采集者紧贴，从而确保标注采集设备输出的6Dof(6自由度，三个自由度描述位置信息，三个自由度描述姿态信息)运动轨迹与人体完全一致。数据采集可以采用多人、多设备类型、多批次、多场景的采集方式，以保证原始惯性测量数据的随机性、丰富性和数据量。本公开实施例对数据采集设备和标注采集设备不做限定，可以根据实际情况进行设置。
步骤5012，分别对各原始惯性测量数据进行姿态转换和坐标转换，得到各原始惯性测量数据对应的样本惯性测量数据。
在训练前，需要对各原始惯性测量数据进行姿态转换，以将各原始惯性测量数据从数据采集姿态的坐标系转换到标准姿态的坐标系；再对各转换后的惯性测量数据进行坐标转换，以将各转换后的惯性测量数据从标准姿态的坐标系转换到世界坐标系下，得到各原始惯性测量数据对应的样本惯性测量数据。在转换到世界坐标系的过程中，可以根据数据采集前校准过程中建立的坐标映射关系，先将标准姿态下的惯性测量数据转换到标注采集设备的坐标系中，再借助标注采集设备的坐标系转换到世界坐标系中。姿态转换和坐标转换过程可以参考上述实施例，本公开实施例在此不再赘述。
步骤5013，将多个样本惯性测量数据的采集时刻与实际轨迹的采集时刻进行时间对准后，确定各样本惯性测量数据对应的实际速度。
由于在校准过程中已经对数据采集设备和标注采集设备进行了时间校准，因此，可以在得到样本惯性测量数据后，根据样本惯性测量数据的采集时刻与实际轨迹的采集时刻进行时间对准，这样，就可以得到各样本惯性测量数据对应的实际速度。
步骤5014，根据多个样本惯性测量数据和各样本惯性测量数据对应的实际速度得到训练样本集。
得到多个样本惯性测量数据和各样本惯性测量数据对应的实际速度后，由多个样本惯性测量数据和各样本惯性测量数据对应的实际速度组成训练样本集。
上述实施例中，分别获取数据采集设备采集的多个原始惯性测量数据和标注采集设备采集的实际轨迹；分别对各原始惯性测量数据进行姿态转换和坐标转换，得到各原始惯性测量数据对应的样本惯性测量数据；将多个样本惯性测量数据的采集时刻与实际轨迹的采集时刻进行时间对准后，确定各样本惯性测量数据对应的实际速度；根据多个样本惯性测 量数据和各样本惯性测量数据对应的实际速度得到训练样本集。通过本公开实施例，在获取样本前，建立了数据采集设备与标注采集设备之间的坐标映射关系，并对两者进行了时间校准，这样，在数据采集后，可以根据坐标映射关系对惯性测量数据进行归一化，而且还可以得到样本惯性测量数据与实际速度之间的对应关系，从而得到训练样本集，为训练速度预测网络做好了前期准备。
应该理解的是，虽然图2至图9的流程图中的各个步骤按照箭头的指示依次显示，但是这些步骤并不是必然按照箭头指示的顺序依次执行。除非本文中有明确的说明，这些步骤的执行并没有严格的顺序限制，这些步骤可以以其它的顺序执行。而且，图2至图9中的至少一部分步骤可以包括多个步骤或者多个阶段，这些步骤或者阶段并不必然是在同一时刻执行完成，而是可以在不同的时刻执行，这些步骤或者阶段的执行顺序也不必然是依次进行，而是可以与其它步骤或者其它步骤中的步骤或者阶段的至少一部分轮流或者交替地执行。
在一个实施例中，如图10所示，提供了一种定位装置，包括：
惯性数据获取模块601，用于获取至少一个用户设备采集的惯性测量数据；
速度预测模块602，用于将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
定位数据确定模块603，用于根据待测终端的速度确定待测终端的定位数据。
在其中一个实施例中，上述惯性数据获取模块601，包括：
原始数据获取子模块，用于获取至少一个用户设备采集的原始惯性测量数据；
转换子模块，用于对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据。
在其中一个实施例中，上述转换子模块，具体用于将原始惯性测量数据转换为用户设备在标准姿态下的第一惯性测量数据；将第一惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在其中一个实施例中，上述转换子模块，具体用于根据原始惯性测量数据获取用户设备的姿态角；根据用户设备的姿态角，将原始惯性测量数据映射至用户设备在标准姿态下的坐标系中，得到第一惯性测量数据。
在其中一个实施例中，上述转换子模块，具体用于根据用户设备在标准姿态下的坐标系与待测终端的坐标系之间的对应关系，将第一惯性测量数据转换至待测终端的坐标系中，得到第二惯性测量数据；根据待测终端的坐标系与世界坐标系之间的对应关系，将第二惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在其中一个实施例中，至少一个用户设备采集的惯性测量数据包括多个用户设备采集的惯性测量数据，如图11所示，该装置还包括：
对准模块604，用于根据各用户设备采集惯性测量数据的时间戳，对各用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据；
上述速度预测模块603，具体用于将对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。
在其中一个实施例中，上述定位数据确定模块603，用于根据待测终端的速度获取待测终端的轨迹信息；将轨迹信息与预设场景地图进行匹配，得到并输出待测终端的位置信息。
在其中一个实施例中，如图12所示，该装置还包括：
服务信息输出模块605，用于根据待测终端的位置信息和预设场景地图，得到并输出服务信息；其中，服务信息包括路径导引信息和物品介绍信息中的至少一种。
在其中一个实施例中，如图13所示，该装置还包括：
样本获取模块606，用于获取训练样本集；其中，训练样本集包括多个样本惯性测量数据和各样本惯性测量数据对应的标注，标注为实际速度；
训练模块607，用于基于训练样本集进行神经网络模型的训练，得到初始预测网络；
测试模块608，用于基于训练样本集对初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹；
网络确定模块609，用于在多个测试轨迹之间的误差小于预设误差值的情况下，将初始预测网络确定为速度预测网络。
在其中一个实施例中，上述样本获取模块606，具体用于分别获取数据采集设备采集的多个原始惯性测量数据和标注采集设备采集的实际轨迹；分别对各原始惯性测量数据进行姿态转换和坐标转换，得到各原始惯性测量数据对应的样本惯性测量数据；将多个样本惯性测量数据的采集时刻与实际轨迹的采集时刻进行时间对准后，确定各样本惯性测量数据对应的实际速度；根据多个样本惯性测量数据和各样本惯性测量数据对应的实际速度得到训练样本集。
在其中一个实施例中，至少一个用户设备包括智能手机、耳机、智能手环、智能手表和智能眼镜中的至少一种。
关于定位装置的具体限定可以参见上文中对于定位方法的限定，在此不再赘述。上述定位装置中的各个模块可全部或部分通过软件、硬件及其组合来实现。上述各模块可以硬件形式内嵌于或独立于终端中的处理器中，也可以以软件形式存储于终端中的存储器中，以便于处理器调用执行以上各个模块对应的操作。
在一个实施例中，提供了一种终端，其内部结构图可以如图14所示。该终端包括通过系统总线连接的处理器、存储器、通信接口、显示屏和输入装置。其中，该终端的处理器用于提供计算和控制能力。该终端的存储器包括非易失性存储介质、内存储器。该非易失性存储介质存储有操作系统和计算机程序。该内存储器为非易失性存储介质中的操作系统和计算机程序的运行提供环境。该终端的通信接口用于与外部的终端进行有线或无线方式的通信，无线方式可通过WIFI、运营商网络、NFC(近场通信)或其他技术实现。该计算机程序被处理器执行时以实现一种定位方法。该终端的显示屏可以是液晶显示屏或者 电子墨水显示屏，该终端的输入装置可以是显示屏上覆盖的触摸层，也可以是终端外壳上设置的按键、轨迹球或触控板，还可以是外接的键盘、触控板或鼠标等。
本领域技术人员可以理解，图14中示出的结构，仅仅是与本申请方案相关的部分结构的框图，并不构成对本申请方案所应用于其上的终端的限定，具体的终端可以包括比图中所示更多或更少的部件，或者组合某些部件，或者具有不同的部件布置。
在一个实施例中，提供了一种终端，包括存储器和处理器，存储器中存储有计算机程序，该处理器执行计算机程序时实现以下步骤：
获取至少一个用户设备采集的惯性测量数据；
将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
根据待测终端的速度确定待测终端的定位数据。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：
获取至少一个用户设备采集的原始惯性测量数据；
对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：
将原始惯性测量数据转换为用户设备在标准姿态下的第一惯性测量数据；
将第一惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：
根据原始惯性测量数据获取用户设备的姿态角；
根据用户设备的姿态角，将原始惯性测量数据映射至用户设备在标准姿态下的坐标系中，得到第一惯性测量数据。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：
根据用户设备在标准姿态下的坐标系与待测终端的坐标系之间的对应关系，将第一惯性测量数据转换至待测终端的坐标系中，得到第二惯性测量数据；
根据待测终端的坐标系与世界坐标系之间的对应关系，将第二惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在一个实施例中，至少一个用户设备采集的惯性测量数据包括多个用户设备采集的惯性测量数据，处理器执行计算机程序时还实现以下步骤：
根据各用户设备采集惯性测量数据的时间戳，对各用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据；
将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度，包括：
将对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：
根据待测终端的速度获取待测终端的轨迹信息；
将轨迹信息与预设场景地图进行匹配，得到并输出待测终端的位置信息。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：
根据待测终端的位置信息和预设场景地图，得到并输出服务信息；其中，服务信息包括路径导引信息和物品介绍信息中的至少一种。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：
获取训练样本集；其中，训练样本集包括多个样本惯性测量数据和各样本惯性测量数据对应的标注，标注为实际速度；
基于训练样本集进行神经网络模型的训练，得到初始预测网络；
基于训练样本集对初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹；
在多个测试轨迹之间的误差小于预设误差值的情况下，将初始预测网络确定为速度预测网络。
在一个实施例中，处理器执行计算机程序时还实现以下步骤：
分别获取数据采集设备采集的多个原始惯性测量数据和标注采集设备采集的实际轨迹；
分别对各原始惯性测量数据进行姿态转换和坐标转换，得到各原始惯性测量数据对应的样本惯性测量数据；
将多个样本惯性测量数据的采集时刻与实际轨迹的采集时刻进行时间对准后，确定各样本惯性测量数据对应的实际速度；
根据多个样本惯性测量数据和各样本惯性测量数据对应的实际速度得到训练样本集。
在一个实施例中，至少一个用户设备包括智能手机、耳机、智能手环、智能手表和智能眼镜中的至少一种。
在一个实施例中，提供了一种计算机可读存储介质，其上存储有计算机程序，计算机程序被处理器执行时实现以下步骤：
获取至少一个用户设备采集的惯性测量数据；
将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
根据待测终端的速度确定待测终端的定位数据。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：
获取至少一个用户设备采集的原始惯性测量数据；
对原始惯性测量数据进行姿态转换和坐标转换，得到惯性测量数据。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：
将原始惯性测量数据转换为用户设备在标准姿态下的第一惯性测量数据；
将第一惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：
根据原始惯性测量数据获取用户设备的姿态角；
根据用户设备的姿态角，将原始惯性测量数据映射至用户设备在标准姿态下的坐标系中，得到第一惯性测量数据。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：
根据用户设备在标准姿态下的坐标系与待测终端的坐标系之间的对应关系，将第一惯性测量数据转换至待测终端的坐标系中，得到第二惯性测量数据；
根据待测终端的坐标系与世界坐标系之间的对应关系，将第二惯性测量数据转换至世界坐标系下，得到惯性测量数据。
在一个实施例中，至少一个用户设备采集的惯性测量数据包括多个用户设备采集的惯性测量数据，计算机程序被处理器执行时还实现以下步骤：
根据各用户设备采集惯性测量数据的时间戳，对各用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据；
将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度，包括：
将对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：
根据待测终端的速度获取待测终端的轨迹信息；
将轨迹信息与预设场景地图进行匹配，得到并输出待测终端的位置信息。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：
根据待测终端的位置信息和预设场景地图，得到并输出服务信息；其中，服务信息包括路径导引信息和物品介绍信息中的至少一种。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：
获取训练样本集；其中，训练样本集包括多个样本惯性测量数据和各样本惯性测量数据对应的标注，标注为实际速度；
基于训练样本集进行神经网络模型的训练，得到初始预测网络；
基于训练样本集对初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹；
在多个测试轨迹之间的误差小于预设误差值的情况下，将初始预测网络确定为速度预测网络。
在一个实施例中，计算机程序被处理器执行时还实现以下步骤：
分别获取数据采集设备采集的多个原始惯性测量数据和标注采集设备采集的实际轨迹；
分别对各原始惯性测量数据进行姿态转换和坐标转换，得到各原始惯性测量数据对应 的样本惯性测量数据；
将多个样本惯性测量数据的采集时刻与实际轨迹的采集时刻进行时间对准后，确定各样本惯性测量数据对应的实际速度；
根据多个样本惯性测量数据和各样本惯性测量数据对应的实际速度得到训练样本集。
在一个实施例中，至少一个用户设备包括智能手机、耳机、智能手环、智能手表和智能眼镜中的至少一种。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流程，是可以通过计算机程序来指令相关的硬件来完成，所述的计算机程序可存储于一非易失性计算机可读取存储介质中，该计算机程序在执行时，可包括如上述各方法的实施例的流程。其中，本申请所提供的各实施例中所使用的对存储器、存储、数据库或其它介质的任何引用，均可包括非易失性和易失性存储器中的至少一种。非易失性存储器可包括只读存储器(Read-Only Memory，ROM)、磁带、软盘、闪存或光存储器等。易失性存储器可包括随机存取存储器(Random Access Memory，RAM)或外部高速缓冲存储器。作为说明而非局限，RAM可以是多种形式，比如静态随机存取存储器(Static Random Access Memory，SRAM)或动态随机存取存储器(Dynamic Random Access Memory，DRAM)等。
以上实施例的各技术特征可以进行任意的组合，为使描述简洁，未对上述实施例中的各个技术特征所有可能的组合都进行描述，然而，只要这些技术特征的组合不存在矛盾，都应当认为是本说明书记载的范围。
以上所述实施例仅表达了本申请的几种实施方式，其描述较为具体和详细，但并不能因此而理解为对发明专利范围的限制。应当指出的是，对于本领域的普通技术人员来说，在不脱离本申请构思的前提下，还可以做出若干变形和改进，这些都属于本申请的保护范围。因此，本申请专利的保护范围应以所附权利要求为准。

Claims (20)

  1. 一种定位方法，其特征在于，所述方法包括：
    获取至少一个用户设备采集的惯性测量数据；
    将所述惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；
    所述速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
    根据所述待测终端的速度确定所述待测终端的定位数据。
  2. 根据权利要求1所述的方法，其特征在于，所述获取至少一个用户设备采集的惯性测量数据，包括：
    获取所述至少一个用户设备采集的原始惯性测量数据；
    对所述原始惯性测量数据进行姿态转换和坐标转换，得到所述惯性测量数据。
  3. 根据权利要求2所述的方法，其特征在于，所述对所述原始惯性测量数据进行姿态转换和坐标转换，得到所述惯性测量数据，包括：
    将所述原始惯性测量数据转换为所述用户设备在标准姿态下的第一惯性测量数据；
    将所述第一惯性测量数据转换至世界坐标系下，得到所述惯性测量数据。
  4. 根据权利要求3所述的方法，其特征在于，所述将所述原始惯性测量数据转换为所述用户设备在标准姿态下的第一惯性测量数据，包括：
    根据所述原始惯性测量数据获取所述用户设备的姿态角；
    根据所述用户设备的姿态角，将所述原始惯性测量数据映射至所述用户设备在标准姿态下的坐标系中，得到所述第一惯性测量数据。
  5. 根据权利要求3所述的方法，其特征在于，所述将所述第一惯性测量数据转换至世界坐标系下，得到所述惯性测量数据，包括：
    根据所述用户设备在标准姿态下的坐标系与所述待测终端的坐标系之间的对应关系，将所述第一惯性测量数据转换至所述待测终端的坐标系中，得到第二惯性测量数据；
    根据所述待测终端的坐标系与所述世界坐标系之间的对应关系，将所述第二惯性测量数据转换至所述世界坐标系下，得到所述惯性测量数据。
  6. 根据权利要求1所述的方法，其特征在于，所述将惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度，包括：
    将多个采集时刻的惯性测量数据中的加速度组成加速度序列，并将所述加速度序列输入到所述速度预测网络中，得到待测终端的速度。
  7. 根据权利要求1-6任一项所述的方法，其特征在于，所述至少一个用户设备采集的惯性测量数据包括多个用户设备采集的惯性测量数据，所述方法还包括：
    根据各所述用户设备采集惯性测量数据的时间戳，对各所述用户设备采集的惯性测量数据进行时间对准，得到对准后的惯性测量数据；
    所述将所述惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度，包括：
    将所述对准后的惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度。
  8. 根据权利要求1所述的方法，其特征在于，所述根据所述待测终端的速度确定所述待测终端的定位数据，包括：
    根据所述待测终端的速度获取所述待测终端的轨迹信息；
    将所述轨迹信息与预设场景地图进行匹配，得到并输出所述待测终端的位置信息。
  9. 根据权利要求8所述的方法，其特征在于，所述方法还包括：
    根据所述待测终端的位置信息和所述预设场景地图，得到并输出服务信息；其中，所述服务信息包括路径导引信息和物品介绍信息中的至少一种。
  10. 根据权利要求9所述的方法，其特征在于，所述方法还包括：
    将所述服务信息与空间音频、语音助手中的至少一种功能结合，以为盲人提供室内语音导航服务。
  11. 根据权利要求1所述的方法，其特征在于，所述速度预测网络的训练过程包括：
    获取训练样本集；其中，所述训练样本集包括多个样本惯性测量数据和各所述样本惯性测量数据对应的标注，所述标注为实际速度；
    基于所述训练样本集进行神经网络模型的训练，得到初始预测网络；
    基于所述训练样本集对所述初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹；
    在所述多个测试轨迹之间的误差小于预设误差值的情况下，将所述初始预测网络确定为所述速度预测网络。
  12. 根据权利要求11所述的方法，其特征在于，所述获取训练样本集，包括：
    分别获取数据采集设备采集的多个原始惯性测量数据和标注采集设备采集的实际轨迹；
    分别对各所述原始惯性测量数据进行姿态转换和坐标转换，得到各所述原始惯性测量数据对应的样本惯性测量数据；
    将多个所述样本惯性测量数据的采集时刻与所述实际轨迹的采集时刻进行时间对准后，确定各所述样本惯性测量数据对应的实际速度；
    根据多个所述样本惯性测量数据和各所述样本惯性测量数据对应的实际速度得到所述训练样本集。
  13. 根据权利要求11所述的方法，其特征在于，所述基于所述训练样本集进行神经网络模型的训练，得到初始预测网络，包括：
    将所述训练样本集中的一个样本惯性测量数据输入到神经网络模型中进行训练，得到所述神经网络模型输出的训练结果；
    利用预设损失函数计算训练结果与所述样本惯性测量数据对应的标注之间的损失值；
    若所述损失值不符合预设收敛条件，则调整神经网络模型中的可调参数，并将另一个样本惯性测量数据输入到参数修改后的神经网络模型中继续进行训练，直到所述神经网络模型 输出的训练结果与对应的标注之间的损失值符合所述预设收敛条件时，结束训练，并将结束训练时的神经网络模型确定为所述初始预测网络。
  14. 根据权利要求11所述的方法，其特征在于，所述方法还包括：
    若多个所述测试轨迹之间的误差大于或等于所述预设误差值，则重复所述基于所述训练样本集对所述初始预测网络进行测试，根据测试得到的预测速度确定多个测试轨迹的步骤，直到多个所述测试轨迹之间的误差小于所述预设误差值时结束训练，得到所述速度预测网络。
  15. 根据权利要求1所述的方法，其特征在于，所述至少一个用户设备包括耳机。
  16. 根据权利要求15所述的方法，其特征在于，所述至少一个用户设备还包括智能手机、智能手环、智能手表和智能眼镜中的至少一种。
  17. 一种定位装置，其特征在于，所述装置包括：
    惯性数据获取模块，用于获取至少一个用户设备采集的惯性测量数据；
    速度预测模块，用于将所述惯性测量数据输入到速度预测网络中进行降噪和速度预测，得到待测终端的速度；所述速度预测网络为采用多个用户设备采集的样本惯性测量数据训练得到；
    定位数据确定模块，用于根据所述待测终端的速度确定所述待测终端的定位数据。
  18. 根据权利要求17所述的装置，其特征在于，所述惯性数据获取模块，包括：
    原始数据获取子模块，用于获取所述至少一个用户设备采集的原始惯性测量数据；
    转换子模块，用于对所述原始惯性测量数据进行姿态转换和坐标转换，得到所述惯性测量数据。
  19. 一种终端，包括存储器和处理器，所述存储器存储有计算机程序，其特征在于，所述处理器执行所述计算机程序时实现权利要求1至16中任一项所述的方法的步骤。
  20. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，所述计算机程序被处理器执行时实现权利要求1至16中任一项所述的方法的步骤。
PCT/CN2022/119859 2021-10-26 2022-09-20 定位方法、装置、终端和存储介质 WO2023071615A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111248781.2A CN113873637A (zh) 2021-10-26 2021-10-26 定位方法、装置、终端和存储介质
CN202111248781.2 2021-10-26

Publications (1)

Publication Number Publication Date
WO2023071615A1 true WO2023071615A1 (zh) 2023-05-04

Family

ID=78998057

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/119859 WO2023071615A1 (zh) 2021-10-26 2022-09-20 定位方法、装置、终端和存储介质

Country Status (2)

Country Link
CN (1) CN113873637A (zh)
WO (1) WO2023071615A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113873637A (zh) * 2021-10-26 2021-12-31 上海瑾盛通信科技有限公司 定位方法、装置、终端和存储介质

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016068742A1 (ru) * 2014-10-28 2016-05-06 Инвенсенс Интернешнл, Инк. Способ и система позиционирования мобильного терминала внутри зданий
CN105823483A (zh) * 2016-05-11 2016-08-03 南京邮电大学 一种基于惯性测量单元的用户步行定位方法
US20190113347A1 (en) * 2017-10-12 2019-04-18 Hanwha Land Systems Co., Ltd. Inertia-based navigation apparatus and inertia-based navigation method based on relative preintegration
CN109883418A (zh) * 2019-01-17 2019-06-14 中国科学院遥感与数字地球研究所 一种室内定位方法及装置
CN110986930A (zh) * 2019-11-29 2020-04-10 北京三快在线科技有限公司 设备定位方法、装置、电子设备及存储介质
CN111007455A (zh) * 2019-10-16 2020-04-14 张苏 定位系统及方法、数据库、神经网络模型训练方法
CN111522034A (zh) * 2020-04-23 2020-08-11 海能达通信股份有限公司 基于惯性导航的定位方法、设备及装置
CN112729301A (zh) * 2020-12-10 2021-04-30 深圳大学 一种基于多源数据融合的室内定位方法
CN113873637A (zh) * 2021-10-26 2021-12-31 上海瑾盛通信科技有限公司 定位方法、装置、终端和存储介质

Family Cites Families (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110533553B (zh) * 2018-05-25 2023-04-07 阿里巴巴集团控股有限公司 服务提供方法和装置
JP7342864B2 (ja) * 2018-07-02 2023-09-12 ソニーグループ株式会社 測位プログラム、測位方法、及び測位装置
CN109068282B (zh) * 2018-09-27 2022-12-02 北京奇虎科技有限公司 一种用于识别用户出行场景的方法、装置及计算机设备
CN113984051A (zh) * 2020-04-30 2022-01-28 深圳市瑞立视多媒体科技有限公司 Imu与刚体的位姿融合方法、装置、设备及存储介质
CN111596298B (zh) * 2020-05-13 2022-10-14 北京百度网讯科技有限公司 目标对象的定位方法、装置、设备及存储介质
CN111397604B (zh) * 2020-06-03 2021-04-27 腾讯科技（深圳）有限公司 一种分析轨迹数据的方法、相关设备及存储介质
CN111881347B (zh) * 2020-07-16 2022-04-15 北京师范大学 基于场景的学习服务推送方法、终端及存储介质
CN112945227A (zh) * 2021-02-01 2021-06-11 北京嘀嘀无限科技发展有限公司 定位方法和装置
CN113252048B (zh) * 2021-05-12 2023-02-28 高新兴物联科技股份有限公司 一种导航定位方法、导航定位系统及计算机可读存储介质

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016068742A1 (ru) * 2014-10-28 2016-05-06 Инвенсенс Интернешнл, Инк. Способ и система позиционирования мобильного терминала внутри зданий
CN105823483A (zh) * 2016-05-11 2016-08-03 南京邮电大学 一种基于惯性测量单元的用户步行定位方法
US20190113347A1 (en) * 2017-10-12 2019-04-18 Hanwha Land Systems Co., Ltd. Inertia-based navigation apparatus and inertia-based navigation method based on relative preintegration
CN109883418A (zh) * 2019-01-17 2019-06-14 中国科学院遥感与数字地球研究所 一种室内定位方法及装置
CN111007455A (zh) * 2019-10-16 2020-04-14 张苏 定位系统及方法、数据库、神经网络模型训练方法
CN110986930A (zh) * 2019-11-29 2020-04-10 北京三快在线科技有限公司 设备定位方法、装置、电子设备及存储介质
CN111522034A (zh) * 2020-04-23 2020-08-11 海能达通信股份有限公司 基于惯性导航的定位方法、设备及装置
CN112729301A (zh) * 2020-12-10 2021-04-30 深圳大学 一种基于多源数据融合的室内定位方法
CN113873637A (zh) * 2021-10-26 2021-12-31 上海瑾盛通信科技有限公司 定位方法、装置、终端和存储介质

Also Published As

Publication number Publication date
CN113873637A (zh) 2021-12-31

Similar Documents

Publication Publication Date Title
US10646139B2 (en) Body movement tracking
JP5987247B2 (ja) データ融合によるモーションキャプチャポインタ
US10837794B2 (en) Method and system for characterization of on foot motion with multiple sensor assemblies
US10652696B2 (en) Method and apparatus for categorizing device use case for on foot motion using motion sensor data
JP2012143342A (ja) 運動解析装置及び運動解析方法
WO2021012879A1 (zh) 轨迹重建方法、装置、计算机设备和存储介质
WO2016131279A1 (zh) 运动轨迹记录方法和用户设备
US20150286279A1 (en) Systems and methods for guiding a user during calibration of a sensor
US20140288873A1 (en) Pre-processing inertial sensor measurements for navigation
WO2023071615A1 (zh) 定位方法、装置、终端和存储介质
CN110044377B (zh) 一种基于Vicon的IMU离线标定方法
CN105068657B (zh) 手势的识别方法及装置
JP2017073753A (ja) 補正方法、プログラム及び電子機器
Roggen et al. An educational and research kit for activity and context recognition from on-body sensors
CN115919250A (zh) 一种人体动态关节角测量系统
US11792517B2 (en) Pose tracking for rolling shutter camera
Young et al. Distributed estimation of linear acceleration for improved accuracy in wireless inertial motion capture
US10748320B2 (en) Animation simulation of biomechanics
WO2023151558A1 (zh) 用于显示图像的方法、装置和电子设备
CN112762915A (zh) 指南针精度优化方法、装置、存储介质及电子设备
CN111382701A (zh) 动作捕捉方法、装置、电子设备及计算机可读存储介质
CN109997014B (zh) 用于确定轨迹的系统和方法
CN104667508A (zh) 方位角校准方法和运动分析装置
CN115844381A (zh) 一种人体动作识别方法、装置、电子设备和存储介质
CN113536552A (zh) 一种人体姿态可视化跟踪系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22885502

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE