WO2017067465A1 - 一种手势识别方法和装置 - Google Patents

一种手势识别方法和装置 Download PDF

Info

Publication number
WO2017067465A1
WO2017067465A1 PCT/CN2016/102603 CN2016102603W WO2017067465A1 WO 2017067465 A1 WO2017067465 A1 WO 2017067465A1 CN 2016102603 W CN2016102603 W CN 2016102603W WO 2017067465 A1 WO2017067465 A1 WO 2017067465A1
Authority
WO
WIPO (PCT)
Prior art keywords
srs signal
signal
terminal device
gesture
srs
Prior art date
Application number
PCT/CN2016/102603
Other languages
English (en)
French (fr)
Inventor
薛剑韬
李安俭
苏滨
Original Assignee
华为技术有限公司
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 华为技术有限公司 filed Critical 华为技术有限公司
Priority to EP16856896.2A priority Critical patent/EP3358495A4/en
Publication of WO2017067465A1 publication Critical patent/WO2017067465A1/zh
Priority to US15/958,380 priority patent/US10732724B2/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
    • G06V40/20—Movements or behaviour, e.g. gesture recognition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/017—Gesture based interaction, e.g. based on a set of recognized hand gestures
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
    • G06F3/0487—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser
    • G06F3/0488—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser using a touch-screen or digitiser, e.g. input of commands through traced gestures
    • G06F3/04883—Interaction techniques based on graphical user interfaces [GUI] using specific features provided by the input device, e.g. functions controlled by the rotation of a mouse with dual sensing arrangements, or of the nature of the input device, e.g. tap gestures based on pressure sensed by a digitiser using a touch-screen or digitiser, e.g. input of commands through traced gestures for inputting data by handwriting, e.g. gesture or text
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0048—Allocation of pilot signals, i.e. of signals known to the receiver
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W72/00—Local resource management
    • H04W72/04—Wireless resource allocation
    • H04W72/044—Wireless resource allocation based on the type of the allocated resource
    • H04W72/0453—Resources in frequency domain, e.g. a carrier in FDMA
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing

Definitions

  • the present application relates to wireless communication technologies, and in particular, to a gesture recognition method and apparatus.
  • gesture recognition technology has been applied to more and more intelligent electronic devices.
  • Smartphones use gesture recognition input, recognize gestures to make calls, open applications, and so on, which can improve the user experience.
  • gesture recognition is relatively mature, and image recognition technology is used to realize the recognition of interactive commands.
  • This image recognition based method generally requires a high-performance camera device, such as Microsoft's Xbox Kinect product. This method acquires static or dynamic images through an imaging device, and then uses computer vision algorithms to analyze images and perform pattern matching to understand the meaning of such gestures and realize gesture recognition.
  • this technology requires a high-performance camera device, and requires a high-performance processor to perform complex image analysis algorithms, which is costly and difficult to miniaturize, and the recognized action needs to face the camera device, and the above disadvantages lead to this. Technology cannot be widely applied.
  • Google Inc. released a demonstration system for smart watches based on gesture recognition control.
  • the user can control the smart watch by using the movement of the finger and the change of the gesture as an input without touching the dial.
  • the user can use the finger to simulate the pullout of the head movement and simulate the winding action in the vicinity of the smart watch to control the menu of the smart watch.
  • Switch operations. Google calls this gesture-operation interaction technology "Project Sol i,” which uses radar, including radio detection and ranging radars. The user's hand posture and the tiny movements of the fingers are measured and converted into information input. The radar transmits radio waves and collects reflections that hit the target.
  • the user's index finger and thumb rub together to simulate a winding action, and the system recognizes the gesture or action and redirects the gesture into the application.
  • this technology requires a relatively expensive millimeter-scale radar microchip to recognize changes in wireless received signals caused by gesture changes, thereby making gesture recognition control costly, which results in only high-end products to support such gesture recognition.
  • gesture recognition function is implemented at low cost.
  • Embodiments of the present invention provide a gesture recognition method and apparatus, which can implement a gesture recognition function at low cost without performing complicated modifications to an existing terminal device.
  • a gesture recognition method is characterized in that, for a terminal device for transmitting a signal of a mobile communication network, the terminal device stores training sample data, and the training sample data includes a reflected signal of a channel sounding reference signal SRS. Correlation spectral feature quantity and corresponding gesture identifier;
  • the method includes,
  • the terminal device matches the correlation spectral feature quantity of the first reflected signal with the training sample data, and identifies the gesture that the gesture sends an object input.
  • the hardware uses the reflected signal of the uplink channel sounding reference signal in the prior art to perform gesture recognition, and does not affect other communication functions of the terminal device when performing gesture recognition, and realizes gesture recognition of the terminal device at a low cost.
  • the training sample data is stored in advance in the terminal device
  • the training sample data is obtained by the terminal device from a network device.
  • the terminal device stores the correlation spectral feature quantity of the second reflected signal and the corresponding gesture identifier as training sample data of the gesture issuing object input gesture into the terminal device.
  • the above method of the first aspect utilizes the reflected signal of the uplink channel sounding reference signal in the prior art to perform gesture training, and stores the training sample data obtained by the gesture training, thereby being used for the gesture recognition operation.
  • the terminal device further sends the correlation spectral feature quantity of the second reflected signal and the corresponding user gesture identifier as training sample data to the network device.
  • the correlation operation includes
  • the reflected signal that exceeds the set distance from the terminal device when reflection occurs is filtered out.
  • the interference of the unrelated signal reflected by the non-gesture can be reduced, and the accuracy of the gesture recognition is greatly improved.
  • the terminal device determines an SRS signal for gesture recognition according to the received information of the SRS signal set.
  • the SRS signal set includes a parameter value interval corresponding to the at least one SRS signal, and the parameter value interval includes any one or more of the following: a transmission interval, an emission bandwidth, a transmission carrier frequency, and a transmission power of the SRS signal;
  • the SRS signal set includes parameter values corresponding to at least one SRS signal, and the parameter values belong to the parameter value interval.
  • the parameter value interval corresponding to the SRS signal includes any one or more of the following: the transmission interval is less than or equal to 10 milliseconds; the transmission bandwidth is greater than or equal to 100 MHz; the transmission carrier frequency is greater than or equal to 10 GHz; and the transmission power is less than or equal to 0 dBm.
  • the second aspect is a terminal device, which is applicable to a mobile communication system, where the terminal device includes a storage unit, a transmitting unit, a receiving unit, and a processing unit, where
  • a storage unit configured to store a database, where the training sample data is stored, where the training sample data includes a correlation spectral feature quantity of the reflected signal of the channel sounding reference signal SRS and a corresponding gesture identifier;
  • a transmitting unit configured to transmit a first SRS signal determined by the processing unit for gesture recognition
  • a receiving unit configured to receive multiple paths of the first SRS signal reflected from the gesture emitting object Reflected signal of the path;
  • a processing unit configured to determine a first SRS signal for gesture recognition, perform correlation operation on the reflected signal received by the receiving unit, and a signal template corresponding to the first SRS signal, to obtain correlation between the first reflected signal
  • the feature quantity is matched, and the correlation spectral feature quantity is matched with the training sample data in the database to identify a gesture in which the gesture emits an object input.
  • the terminal device is not required to be complicatedly modified or other additional hardware is added, and the reflected signal of the uplink channel sounding reference signal in the prior art is used for gesture recognition, and the gesture recognition is not affected.
  • the other communication functions of the terminal device realize the gesture recognition function of the terminal device at a low cost.
  • the training sample data is stored in advance in the storage unit
  • the training sample data is obtained by the terminal device from the network device by the receiving unit, and stored in a database of the storage unit.
  • the transmitting unit is further configured to transmit a second SRS signal that is determined by the processing unit for gesture recognition;
  • the receiving unit is further configured to receive a second reflected signal of the multiple paths reflected by the second SRS signal from the gesture emitting object;
  • the processing unit is configured to determine a second SRS signal for gesture recognition, and perform a correlation operation between the second reflected signal received by the receiving unit and a signal template corresponding to the second SRS signal to obtain the A correlation spectral feature quantity of the second reflected signal, and storing the correlation spectral feature quantity and the corresponding user gesture identifier as training sample data of the gesture issuing object input gesture into the storage unit.
  • the device of the second aspect uses the reflected signal of the uplink channel sounding reference signal in the prior art to perform gesture training, and stores the training sample data obtained by the gesture training, so that the device can be used. For gesture recognition operations.
  • the transmitting unit further sends the correlation spectral feature quantity of the second reflected signal and the corresponding user gesture identifier as training sample data to the network device.
  • the processing unit row correlation operation including,
  • the reflected signal that exceeds the set distance from the terminal device when reflection occurs is filtered out.
  • the interference of the unrelated signal reflected by the non-gesture can be reduced, and the accuracy of the gesture recognition is greatly improved.
  • the storage unit stores an SRS signal set in advance.
  • the processing unit determines an SRS signal for gesture recognition according to the SRS signal set
  • the processing unit determines an SRS signal for gesture recognition according to information of the SRS signal set received by the receiving unit.
  • the SRS signal set includes a parameter value interval corresponding to the at least one SRS signal, and the parameter value interval includes any one or more of the following: a transmission interval, an emission bandwidth, a transmission carrier frequency, and a transmission power of the SRS signal;
  • the SRS signal set includes parameter values corresponding to at least one SRS signal, and the parameter values belong to the parameter value interval.
  • the parameter value interval corresponding to the SRS signal includes any one or more of the following: the transmission interval is less than or equal to 10 milliseconds; the transmission bandwidth is greater than or equal to 100 MHz; and the transmission carrier frequency is greater than or equal to 10 GHz; the transmission power is less than or equal to 0 dBm.
  • a third aspect is a method of generating a channel sounding reference signal SRS for gesture recognition, characterized in that
  • the network device generates an SRS signal set for gesture recognition
  • the network device sends the information about the SRS signal set for gesture recognition to the terminal device;
  • the SRS signal set includes a parameter value interval corresponding to the at least one SRS signal, and the parameter value interval includes any one or more of the following: the transmission interval is less than or equal to 10 milliseconds; the transmission bandwidth is greater than or equal to 100 MHz; and the transmission carrier frequency is greater than or equal to 10 GHz; the transmission power is less than or equal to 0 dBm;
  • the SRS signal set includes parameter values corresponding to at least one SRS signal, and the parameter values belong to the parameter value interval.
  • the fourth aspect a network device, which is applicable to a mobile communication system, where the network device includes a transmitting unit and a processing unit, where
  • the processing unit generates a SRS signal set for gesture recognition, and transmits information of the SRS signal set to the transmitting unit;
  • the transmitting unit is configured to send information about the SRS signal set generated by the processing unit to a terminal device;
  • the SRS signal set includes a parameter value interval corresponding to the at least one SRS signal, and the parameter value interval includes any one or more of the following: the transmission interval is less than or equal to 10 milliseconds; the transmission bandwidth is greater than or equal to 100 MHz; and the transmission carrier frequency is greater than or equal to 10 GHz; the transmission power is less than or equal to 0 dBm;
  • the SRS signal set includes parameter values corresponding to at least one SRS signal, and the parameter values belong to the parameter value interval.
  • FIG. 1 is a flowchart of a SRS signal set configuration according to an embodiment of the present invention
  • FIG. 2 is a flowchart of training of a gesture recognition method according to an embodiment of the present invention
  • FIG. 3 is a schematic structural diagram of data of a training sample according to an embodiment of the present invention.
  • FIG. 4 is a flowchart of a gesture recognition method according to an embodiment of the present invention.
  • FIG. 5 is a flowchart of processing after a gesture recognition method fails to be recognized according to an embodiment of the present invention
  • FIG. 6 is a schematic structural diagram of a terminal device according to an embodiment of the present disclosure.
  • FIG. 7 is a schematic structural diagram of a network device according to an embodiment of the present invention.
  • the concept of the gesture recognition method disclosed in the embodiment of the present application is that the terminal device transmits an uplink channel sounding reference signal (Sound Signal), which is reflected from the user's hand or other object used for gestures back to the terminal device.
  • the terminal device performs gesture recognition on at least one of a power spectrum waveform of the reflected signal of the SRS signal, and a measured amount of arrival time, intensity, phase, and the like of the reflected signal as a feature amount.
  • the network elements involved in the embodiments of the present application are: a network device and a terminal device.
  • the access network device is an eNB, and the core network device is an MME; in the UMTS system, the access network device is an RNC, and the core network device is an SGSN; in other wireless communication systems, there is also a corresponding access network.
  • Equipment and core network equipment In the embodiment of the present application, the foregoing access network device and the core network device are collectively referred to as a network device with respect to the user equipment device.
  • a terminal device in the embodiment of the present disclosure may be a user terminal, a user equipment, and a mobile station.
  • the terminal device determines and transmits an SRS signal for gesture recognition, receives a reflected signal of the SRS signal for a user gesture, and performs processing; and is configured to be stored on the terminal device side.
  • a database of training sample data The following is introduced separately.
  • the SRS signal is mainly used for channel quality estimation, thereby enabling the base station to perform frequency selective scheduling for the uplink, including power control enhancement and various initialization modulation and coding scheme selection, timing advance, and the like of the terminal device.
  • the SRS signal sent by the terminal device is configured by a network device specific broadcast signaling indication.
  • the terminal device configures the SRS signal according to the network device indication and transmits it at the specified location.
  • the time-frequency resources occupied by the SRS signal such as the transmission interval, the transmission bandwidth, the transmission carrier frequency, and the transmission power, are all configured by the network device.
  • the embodiment of the present invention provides a new gesture recognition function for the SRS signal.
  • the network device needs to configure the SRS signal set for gesture recognition for the terminal device.
  • the transmission interval of the SRS signal for gesture recognition is smaller than that of the conventional SRS signal for channel quality estimation; however, the transmission interval of the SRS signal is set too small, and the SRS signal is transmitted too frequently. It can cause interference to other communication signals, and at the same time poses a huge challenge to the computing power of the terminal device.
  • the transmission interval of the SRS signal for gesture recognition in the embodiment of the present invention may be set to be less than 10 milliseconds, and the transmission frequency may be sufficient to identify the motion gesture of the user.
  • the specific value may be 5 milliseconds or 10 milliseconds. Wait.
  • the SRS signal for gesture recognition has a larger transmission bandwidth and carrier frequency than the conventional SRS signal used for channel quality estimation.
  • the transmission bandwidth of the SRS signal for gesture recognition is set to 100 MHz or more, so that the terminal device judges The distance accuracy can reach 30 cm; the carrier frequency of the SRS signal for gesture recognition is set to be greater than 10 GHz, so that the terminal device can determine that the gesture change action can reach 1 cm.
  • the SRS signal for gesture recognition acts only near the terminal device, and does not need to be transmitted to the network device, and the transmit power is appropriately reduced in order to save energy and avoid interference with normal wireless communication.
  • the SRS signal transmission power used for channel quality estimation in the prior art is 23 dBm
  • the SRS signal transmission power for gesture recognition in the embodiment of the present application may be reduced to 0 dBm.
  • step 101 the network device generates an SRS signal set for gesture recognition.
  • the SRS signal set includes a parameter value interval corresponding to the at least one SRS signal, and the parameter value interval includes any one or more of the following: a transmission interval, an emission bandwidth, a transmission carrier frequency, and a transmission power of the SRS signal.
  • the SRS signal set can be set such that the transmission interval is less than or equal to 10 milliseconds; the transmission bandwidth is greater than or equal to 100 MHz; the transmission carrier frequency is greater than or equal to 10 GHz; and the transmission power is less than or equal to 0 dBm.
  • the terminal device can determine, according to the SRS signal set for gesture recognition, that the parameter corresponding to the SRS signal that needs to be transmitted currently is: a transmission interval of 5 milliseconds, a transmission bandwidth of 100 MHz, a transmission carrier frequency of 10 GHz, and a transmission power. It is 0 dBm; it can also be determined as: the transmission interval is 10 milliseconds, the transmission bandwidth is 150 MHz, the transmission carrier frequency is 12 GHz, and the transmission power is 0 dBm, etc., as long as the requirements of the SRS signal set are satisfied.
  • the signal set may further include a parameter value corresponding to a specific SRS signal.
  • the SRS signal set may be set such that the SRS signal 1 has a transmission interval of 5 milliseconds, an emission bandwidth of 100 MHz, and a transmission carrier frequency of 10 GHz. And the transmission power is 0 dBm; the SRS signal 2, the transmission interval is 10 milliseconds, the transmission bandwidth is 150 MHz, the transmission carrier frequency is 12 GHz, and the transmission power is 0 dBm; etc., as long as the requirements of the SRS signal set are satisfied.
  • Step 102 The network device sends the information about the SRS signal set for gesture recognition to the terminal device.
  • Step 103 The terminal device receives the information of the SRS signal set, and according to the SRS signal set.
  • the information determines the SRS signal currently used for gesture recognition.
  • the terminal device selects an appropriate SRS signal from the SRS signal set for gesture recognition for transmission and identification.
  • the SRS signal set for gesture recognition specifies a parameter value interval of the SRS signal
  • the terminal device generates an SRS signal that meets the above-mentioned parameter value interval constraint;
  • the SRS signal set for gesture recognition specifies a specific set of SRS signals
  • the terminal device selects one of the specific SRS signal parameter values. Eligible SRS signals can be selected according to the specific application environment and needs.
  • the SRS signal set for gesture recognition can be solidified in the terminal device.
  • the SRS signal set configuration process can be omitted, and the terminal device does not need the network device to generate and configure the SRS signal set.
  • the SRS signal is directly selected from the solidified SRS signal set for gesture recognition for transmission.
  • the terminal device may first go through a training phase, record the spectral feature quantity of the reflected signal of the SRS signal of the user's certain gesture, and store it as training sample data in the database.
  • step 201 the terminal device transmits an SRS signal.
  • the terminal device For the SRS signal sent by the terminal device, if there is a SRS signal set configuration process of the network device, the terminal device selects the SRS signal from the SRS signal sent by the network device; if there is no SRS signal set configuration process of the network device, Then, the terminal device selects the SRS signal from the self-cured SRS signal set for gesture recognition.
  • the manner of selecting a specific SRS signal is similar to that of step 103 and will not be described again.
  • Step 202 The user makes a gesture by a certain distance from the terminal device.
  • the user can gesture near the terminal device, for example, within 30 cm.
  • the gestures mentioned in the embodiments of the present application are not limited to the gestures that are generated by the user's hands, and may be gestures that the user uses other objects to play, or may be gestures played by the animals.
  • Step 203 The terminal device receives the reflected signal of multiple paths for transmitting the SRS signal in step 201. And perform correlation operation with the signal template corresponding to the SRS signal transmitted in step 201.
  • the terminal device after the terminal device transmits the SRS signal in the uplink in step 201, it is required to immediately detect the reflected signal of the SRS signal at the same frequency, which is a function that the terminal device does not currently have.
  • the prior art terminal device detects only the downlink of the base station, and does not detect the uplink of itself or other terminal devices. Therefore, in order to prevent the reflected signal of the SRS signal from being overwhelmed by the signal transmitted by the terminal device itself, full-duplex technology is required to reduce or eliminate the influence of the SRS signal transmitted by the terminal device itself on the detection of the reflected signal of the SRS signal. For example, the terminal device can subtract the transmitted signal from the received signal, thereby eliminating the influence of the transmitted signal on the received signal in the same channel at the same time.
  • the terminal device can also eliminate the influence of the transmitted signal on the received signal by using a local correlation method, and perform correlation operation on the received signal template of the SRS signal of the multiple paths and the signal template corresponding to the transmitted SRS signal to obtain a correlation spectrum, and obtain a correlation spectrum.
  • the correlated spectral signature of the reflected signal can be obtained, including the phase, power and delay of the reflected signal of the SRS signal for each reflected path.
  • the signal template corresponding to the transmitted SRS signal may be the transmitted SRS signal itself. Power and time delay reflect information such as the size, distance, and the like of the reflector, and the terminal device can be used to recognize different gestures, especially still gestures. Due to the Doppler effect, the phase changes reflect the moving speed and direction of movement of the reflector, and the terminal device can also be used to identify different gestures, especially moving gestures.
  • Step 204 filtering out the reflected signal of the SRS signal exceeding the set distance.
  • the terminal device Since in step 203, the terminal device performs a correlation operation on the received signal signals of the plurality of path SRS signals and the signal template corresponding to the transmitted SRS signal, the arrival time of the reflected signal of each path SRS signal can be calculated, and thus The distance from the terminal device when the SRS is transmitted is calculated.
  • the terminal device can filter out the reflected signal of the SRS signal that is far away from the terminal device when the transmission occurs.
  • the reflected signal of the filtered SRS signal is far away from the SRS signal reflected by the user's hand due to the long distance of the transmitted signal, so that the interference of the unrelated signal can be reduced after filtering, and the accuracy of the gesture recognition is greatly improved. For example, you can filter A reflected signal of the SRS signal other than 30 cm.
  • the step 204 is an optional step and can be omitted in practical applications.
  • Step 205 Record a correlation spectral feature quantity of the reflected signal of the filtered SRS signal, and store the correlation spectral feature quantity and the corresponding user gesture identifier as training sample data in a database.
  • the training sample data 301 includes: a correlation spectral feature amount 303, and a user gesture identifier 302 corresponding to the related spectral feature amount.
  • a correlation spectral feature amount 303 There is only one training sample data of the same gesture, but there may be multiple correlation spectral feature quantities in the gesture training sample data, and multiple correlation spectral feature quantities may be obtained by repeating steps 201-206.
  • the plurality of correlation spectral feature quantities may be processed by a weighted average algorithm, and the more relevant spectral feature quantities, the higher the accuracy in the subsequent gesture recognition phase.
  • a plurality of training sample data as described in FIG. 3 may be stored in the database, corresponding to a plurality of user gestures, respectively.
  • the storage location of the database it may be a non-volatile memory in the terminal device or a non-volatile memory in the network device. If stored in a network device, the terminal device needs to transmit the training sample data to the network device using its supported communication link.
  • the training sample data can be shared among all terminal devices, and the training sample data can be further processed by using a big data analysis method to improve the identification accuracy. Sex.
  • the training sample data of each gesture is obtained in the database; or the training sample data of each gesture is obtained from the database of the network device by sharing; the training sample data It may also be pre-cured in the terminal device database.
  • the terminal device compares the correlation spectral feature amount of the user gesture with the correlation spectral feature amount in the training sample data in the database, thereby identifying the user's gesture.
  • Step 401 In the same step as the training phase, the terminal device transmits an SRS signal.
  • Step 402 in the same step as the training phase step 202, the user is at a certain distance from the terminal device, and plays gesture.
  • Step 403 In the same step 203 as the training phase, the terminal device receives the reflected signals of the multiple paths of the SRS signal transmitted in step 401, and performs a correlation operation with the signal template corresponding to the SRS signal sent in step 401.
  • Step 404 with the training phase step 204, filtering out the reflected signal of the SRS signal exceeding the set distance.
  • Step 405 Record the correlation spectral feature quantity of the reflected signal of the filtered SRS signal, match the measured feature quantity with the training sample data in the database, and identify the input gesture.
  • the reflected signal arrival time, the reflected power, the reflected signal phase change, and the reflected signal path number can be combined to determine the gesture, and the influence of the environment on the gesture recognition is eliminated.
  • the mobile terminal compares the correlation spectral feature quantity of the current record with the correlation spectral feature quantity of each gesture training sample data in the database, and searches for the closest correlation with the current correlation spectral feature quantity.
  • the training sample data is used to determine whether it is the type of user gesture of this time.
  • the difference accumulated value of each feature quantity may be used to determine that the smallest difference accumulated value is the closest training sample data; if the difference accumulated value is less than the set threshold
  • the user gesture type corresponding to the training sample is considered to be the user gesture identified this time.
  • other mathematical algorithms can also be used for judgment, and will not be described herein.
  • the identification may be caused by the surrounding environment interference, and the identification may be repeated.
  • the operation as shown in FIG. 5 can also be performed:
  • Step 501 Determine that the user gesture type corresponding to the current spectral feature quantity is not retrieved from the database, and the recognition fails;
  • Step 502 prompting that the recognition gesture fails
  • Step 503 determining whether the gesture has been trained.
  • Step 504 if it is determined in step 503 that the gesture has been trained, the acquisition is performed.
  • the user gesture is identified, and step 505 is performed;
  • Step 505 Add the correlation spectral feature quantity as a new correlation spectral feature quantity of the input gesture training sample data to the database;
  • Step 506 if it is determined in step 503 that the gesture has not been trained, optionally, prompt whether to newly establish training sample data for the current gesture, and perform step 507;
  • Step 507 If it is confirmed that the training sample data is newly established, and the user gesture identifier of the newly established training sample data is determined, the related spectral feature quantity is stored as the first related spectral feature quantity of the current gesture together with the user gesture identifier. In the database.
  • the training sample data obtained by the gesture training is performed by performing the gesture training and recognition by using the reflected signal of the uplink channel sounding reference signal in the prior art without complicated modification or adding other additional hardware.
  • Storage which can be used for gesture recognition operations, and does not affect other communication functions of the terminal device when performing gesture recognition, and realizes the gesture recognition function of the terminal device at a lower cost; in addition, when processing the reflected signal, The reflected signal exceeding the set distance is filtered to reduce the interference of the unrelated signal reflected by the non-gesture, and the accuracy of the gesture recognition is greatly improved.
  • Embodiment 2 corresponds to Embodiment 1, and is a terminal device and a network device that perform the gesture recognition method described in Embodiment 1.
  • Embodiment 2 of the present invention will be described below with reference to FIG. 6 and FIG.
  • the terminal device 601 includes the following components: a processing unit 602, a transmitting unit 603, a receiving unit 604, a storage unit 605, an input unit 606, and a display unit 607.
  • the processing unit 602 is configured to perform a control operation and a mathematical operation operation of each component unit of the terminal device; the transmitting unit 603 is configured to send an electromagnetic wave signal to the outside of the terminal device; and the receiving unit 604 is configured to receive an electromagnetic wave signal of the external device;
  • the storage unit 605 is configured to store data, a program, and the like; the input unit 606 is configured to receive an input letter of the user of the terminal device.
  • the display unit 607 is configured to display information to a user of the terminal device.
  • the network device 701 includes: a processing unit 702, a transmitting unit 703, a receiving unit 704, and a storage unit 705.
  • terminal device 601 and the network device 701 should also include other necessary components for completing their communication functions, such as a power supply unit, etc., but since the above components are not related to the present invention, they are not described herein again.
  • the configuration information of the SRS signal is the same as that of the first embodiment, and details are not described herein again.
  • Processing unit 702 of network device 701 generates a set of SRS signals for gesture recognition.
  • the content and configuration of the signal set are the same as those in step 101 of Embodiment 1, and are not described again.
  • the network device 701 processing unit 702 transmits the generated SRS signal set for gesture recognition to the transmitting unit 703, and the transmitting unit 703 transmits the SRS signal set information to the terminal device 601.
  • the receiving unit 604 of the terminal device 601 receives the SRS signal set information for gesture recognition, and transmits the information to the processing unit 602.
  • the processing unit 602 determines the current for gesture recognition according to the SRS signal set information received by the receiving unit 604. SRS signal.
  • the processing unit 602 of the terminal device 601 selects an appropriate SRS signal from the SRS signal set for gesture recognition for transmission and identification.
  • the specific selection method is the same as step 103 of the first embodiment.
  • the SRS signal set for gesture recognition may be solidified in the storage unit 605 of the terminal device 601.
  • the SRS signal set configuration process may be omitted, and the terminal device 601 does not need the network device 701 to generate and configure the SRS signal. set.
  • the SRS signal is directly collected from the SRS signal for gesture recognition solidified in the storage unit 605 to be transmitted.
  • the terminal device 601 may first go through a training phase, and record the spectral feature quantity of the reflected signal of the SRS signal of the user's gesture, and store it as training sample data. Stored in the database.
  • the transmitting unit 603 of the terminal device 601 transmits an SRS signal.
  • the terminal device 601 selects the SRS signal from the SRS signal set sent by the network device 701; if there is no SRS of the network device 701 In the signal set configuration process, the terminal device 601 selects the SRS signal from the SRS signal set for gesture recognition that is solidified by the self storage unit 605.
  • the user makes a gesture by a certain distance from the terminal device 601.
  • the user can gesture near the terminal device 601, for example, within 30 cm.
  • the receiving unit 604 of the terminal device 601 receives the reflected signal of the SRS signal of the multiple paths of the SRS signal, and the processing unit 602 transmits the reflected signal of the SRS signal and the signal corresponding to the transmitted SRS signal.
  • the template performs related operations.
  • the method for canceling the influence of the SRS signal sent by the terminal device itself on the detection of the reflected signal of the SRS signal is the same as the step 203 of the first embodiment, and is not described here.
  • the processing unit 602 of the terminal device 601 filters out the reflected signal of the SRS signal exceeding the set distance.
  • the manner in which the signal is filtered out in Embodiment 2 of the present application is the same as that in Step 1 of Embodiment 1, and details are not described herein again.
  • the processing unit 602 of the terminal device 601 records the correlation spectral feature quantity of the reflected signal of the filtered SRS signal, and stores the correlation spectral feature quantity and the corresponding user gesture identifier as the training sample data in the storage unit 605. .
  • the composition of the training sample data is the same as that of step 205 of Embodiment 1, and details are not described herein again.
  • the storage location of the training sample data may be the storage unit 605 in the terminal device 601 or the storage unit in the network device 701. If stored in the network device 701, the processing unit of the terminal device 601 needs to transmit the training sample data to the transmitting unit 603 and to the network device 701. The receiving unit 704 of the network device 701 receives the above training samples And transmitted to the processing unit 702, the processing unit 702 stores the training sample data in the storage unit 705.
  • the training sample data can be shared among all terminal devices, and the training sample data can be further analyzed using a method of big data analysis. Process to improve recognition accuracy.
  • the storage unit 605 has the training sample data of each gesture; or the training sample of each gesture is obtained from the storage unit 705 of the network device 701 in a shared manner.
  • the training sample data may also be pre-cured in the storage unit 605 of the terminal device 601.
  • the processing unit 602 of the terminal device 601 compares the correlation spectral feature amount of the user gesture with the correlation spectral feature amount in the training sample data in the storage unit 605, thereby identifying the user's gesture.
  • the transmitting unit of the terminal device 601 transmits an SRS signal.
  • step 402 of Embodiment 1 the user makes a gesture by a certain distance from the terminal device 601.
  • the receiving unit 604 of the terminal device 601 receives the reflected signals of the plurality of paths for transmitting the SRS signals, and the processing unit 602 performs correlation operations on the signal templates corresponding to the transmitted SRS signals.
  • the processing unit 602 of the terminal device 601 filters out the reflected signal of the SRS signal exceeding the set distance.
  • the processing unit 602 of the terminal device 601 records the correlation spectral feature quantity of the reflected signal of the filtered SRS signal, and matches the measured correlation spectral feature quantity with the training sample data in the storage unit 605. Identify input gestures.
  • the manner in which the processing unit 602 recognizes the input gesture is the same as the step 405 in Embodiment 1, and details are not described herein again.
  • the identification may be caused by the surrounding environment interference, and the identification may be repeated. In addition, it can also be attached as attached The operation described in Figure 5:
  • Step 501 the processing unit 602 determines that the user gesture type corresponding to the current correlation spectral feature quantity is not retrieved from the storage unit 605, and the recognition fails;
  • Step 502 the processing unit 602 controls the display unit 607 to prompt the recognition gesture to fail;
  • Step 503 the processing unit 602 determines, according to the input of the input unit 606, whether the gesture has been trained.
  • Step 504 if it is determined in step 503 that the gesture has been trained, the processing unit 602 determines the current user gesture identifier according to the input of the input unit 606, and performs step 505;
  • Step 505 the processing unit 602 adds the current correlation spectral feature quantity as a new correlation spectral feature quantity of the input gesture training sample data to the storage unit 605;
  • Step 506 if it is determined in step 503 that the gesture has not been trained, optionally, the processing unit 602 controls the display unit 607 to prompt whether to newly establish training sample data for the current gesture, and step 507 is performed;
  • Step 507 if the processing unit 602 confirms the newly established training sample data according to the input of the input unit 606, the processing unit 602 confirms the user gesture identifier of the newly established training sample data according to the input of the input unit 606, and uses the current correlation spectral feature quantity as The first related spectral feature amount of the gesture is stored in the storage unit 605 along with the user gesture identification.
  • Embodiment 2 The beneficial technical effects obtained in Embodiment 2 are the same as those in Embodiment 1, and will not be described again.
  • the embodiment of the present application introduces the technical solution by using the user gesture recognition as an example, those skilled in the art should know that the technical solution of the embodiment of the present application is not limited to the recognition of the user gesture, but also can be used for the shape of other devices. Action recognition.
  • the functional units in the various embodiments of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.
  • the integrated unit is implemented in the form of a software functional unit and sold as a separate product When sold or used, it can be stored on a computer readable storage medium.
  • the technical solution of the present invention contributes in essence or to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium.
  • a number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention.
  • the foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, and the like. .
  • the disclosed system, apparatus, and method may be implemented in other manners, and different implementations may be implemented without any contradiction.
  • the way can be combined.
  • the device embodiments described above are merely illustrative.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or Can be integrated into another system, or some features can be ignored or not executed.
  • the mutual coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, or an electrical, mechanical or other form of connection.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Human Computer Interaction (AREA)
  • Signal Processing (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Data Mining & Analysis (AREA)
  • Psychiatry (AREA)
  • Social Psychology (AREA)
  • Health & Medical Sciences (AREA)
  • Multimedia (AREA)
  • General Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Artificial Intelligence (AREA)
  • Telephone Function (AREA)
  • User Interface Of Digital Computer (AREA)
  • Telephonic Communication Services (AREA)

Abstract

一种手势识别的方法和装置，其中方法用于基于移动通信网络信号的终端设备，该终端设备中存储训练样本数据，训练样本数据包括信道探测参考信号SRS的反射信号的相关谱特征量以及对应的用户手势标识；该方法包括，终端设备发射第一信道探测参考信号SRS；终端设备接收第一信道探测参考信号SRS从手势发出物体反射的多个路径的反射信号，并与第一信道探测参考信号SRS所对应的信号模板进行相关运算，得到反射信号的相关谱特征量；该终端设备将反射信号的相关谱特征量与训练样本数据进行匹配，识别输入手势。

Description

一种手势识别方法和装置
本申请要求于2015年10月21日提交中国专利局、申请号为201510694748.0、发明名称为“一种手势识别方法和装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及无线通信技术，尤其涉及一种手势识别方法和装置。
背景技术
随着智能电子设备的增多，人机交互技术的发展，手势识别技术被应用到越来越多的智能电子设备中。智能手机利用手势识别输入，识别手势进行拨打电话、打开应用程序等操作，能够很好地提升用户体验。
现有技术中手势识别比较成熟的是利用图像识别技术实现交互命令的识别。这种基于图像识别的方法一般需要配备高性能的摄像装置，比较典型的应用如微软的Xbox Kinect产品。这种方法通过摄像设备获取静态或动态图像，然后利用计算机视觉算法分析图像，进行模式匹配，从而理解这种手势的含义，实现手势识别。但是，这一技术需要高性能的摄像装置，并且需要高性能的处理器进行复杂的图像分析算法，成本高昂且难以小型化，而且被识别的动作需要面对摄像装置，以上缺点都导致这一技术无法广泛应用。
近期，谷歌公司发布了基于手势识别控制的智能手表的演示系统。用户无需触碰表盘，就可以利用手指的移动和手势的变化作为输入来控制智能手表，例如，用户可以在智能手表附近用手指模拟拔出表头动作、模拟上弦动作，来控制智能手表的菜单切换操作。谷歌将这种手势操作交互技术称为“Project Sol i”，它使用雷达，包括无线电探测和测距雷达，来探 测用户手部姿势和手指的微小动作，并把它们转化进行信息输入。雷达传输无线电波并收集击中目标的反射感应。比如，用户的食指和大拇指在一起摩擦模拟上弦动作，系统识别这种手势或动作并把手势重定向输入到应用程序中。但是，这一技术需要比较昂贵的毫米级雷达微芯片来识别手势变化导致的无线接收信号的改变，从而进行手势识别控制，成本高昂，这就导致只有高端产品才能支持这类手势识别。
综上所述，由于现有手势识别技术都需要在设备上增加新的高成本组件，导致成本都高昂，本领域亟需一种新型的手势识别方法和技术，尽量减少对用户设备的硬件改动，低成本地实现手势识别功能。
发明内容
本发明的实施例提供了一种手势识别方法和装置，能够在对现有终端设备不进行复杂改动的前提下，低成本地实现手势识别功能。
第一方面，一种手势识别方法，其特征在于，用于传输移动通信网络信号的终端设备，所述终端设备中存储训练样本数据，所述训练样本数据包括信道探测参考信号SRS的反射信号的相关谱特征量以及对应的手势标识；
所述方法包括，
所述终端设备发射第一SRS信号；
所述终端设备接收所述第一SRS信号从手势发出物体反射的多个路径的第一反射信号，并与所述第一SRS信号所对应的信号模板进行相关运算，得到所述第一反射信号的相关谱特征量；
所述终端设备将所述第一反射信号的相关谱特征量与所述训练样本数据进行匹配，识别所述手势发出物体输入的手势。
根据所述第一方面，无需对终端设备进行复杂的改进或增加其他额外 硬件，利用现有技术中上行信道探测参考信号的反射信号来进行手势识别，而且在进行手势识别时也不会影响到终端设备的其他通信功能，以较低的成本实现了终端设备的手势识别功能。
可选的，结合第一方面，对于所述方法，
所述训练样本数据为事先存储在终端设备中的；
或者，
所述训练样本数据为所述终端设备从网络设备得到的。
可选的，结合第一方面，对于所述方法，
所述终端设备发射第二SRS信号；
所述终端设备接收所述第二SRS信号从手势发出物体反射的多个路径的第二反射信号，并与所述第二SRS信号所对应的信号模板进行相关运算，得到所述第二反射信号的相关谱特征量；
所述终端设备将所述第二反射信号的相关谱特征量以及对应的手势标识作为所述手势发出物体输入手势的训练样本数据存储到所述终端设备中。
第一方面的上述方法，利用现有技术中上行信道探测参考信号的反射信号来进行手势训练，将手势训练得到的训练样本数据存储，从而可以用于手势识别操作。
可选的，结合第一方面，对于所述方法，
所述终端设备进一步将所述第二反射信号的相关谱特征量以及对应的用户手势标识作为训练样本数据发送到网络设备中。
可选的，结合第一方面，对于所述方法，
所述相关运算包括，
滤掉发生反射时距离所述终端设备超过设定距离的反射信号。
第一方面的上述方法，过滤超过设定距离的反射信号后，可以减少非手势所反射的无关信号的干扰，大大提高手势识别的精确度。
可选的，结合第一方面，对于所述方法，
所述终端设备根据事先存储在终端设备中的SRS信号集确定用于手势识别的SRS信号；
或者，
所述终端设备接收网络设备发送的SRS信号集的信息；
所述终端设备根据接收到的SRS信号集的信息确定用于手势识别的SRS信号。
可选的，结合第一方面，对于所述方法，
所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：SRS信号的发射间隔、发射带宽、发射载频和发射功率；
或者，
所述SRS信号集包括至少一个SRS信号所对应的参数值，所述参数值属于所述参数值区间。
可选的，结合第一方面，对于所述方法，
所述SRS信号所对应的参数值区间包括下述任意一项或者多项：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于10GHz；发射功率小于等于0dBm。
第二方面，一种终端设备，其特征在于，适用于移动通信系统，所述终端设备包括，存储单元、发射单元、接收单元及处理单元，其中，
存储单元，用于存储数据库，所述数据库中存储有训练样本数据，所述训练样本数据包括信道探测参考信号SRS的反射信号的相关谱特征量以及对应的手势标识；
发射单元，用于发射所述处理单元确定的用于手势识别的第一SRS信号；
接收单元，用于接收所述第一SRS信号从手势发出物体反射的多个路 径的反射信号；
处理单元，用于确定用于手势识别的第一SRS信号，将接收单元接收的所述反射信号与所述第一SRS信号所对应的信号模板进行相关运算，得到所述第一反射信号的相关谱特征量，并将所述相关谱特征量与数据库中的训练样本数据进行匹配，识别所述手势发出物体输入的手势。
根据所述第二方面，无需对终端设备进行复杂的改进或增加其他额外硬件，利用现有技术中上行信道探测参考信号的反射信号来进行手势识别，而且在进行手势识别时也不会影响到终端设备的其他通信功能，以较低的成本实现了终端设备的手势识别功能。
可选的，结合第二方面，对于所述装置，
所述训练样本数据为事先存储在所述存储单元的；
或者，
所述训练样本数据为所述终端设备通过所述接收单元从网络设备得到，并存入所述所述存储单元的数据库。
可选的，结合第二方面，对于所述装置，
所述发射单元，进一步用于发射所述处理单元确定的用于手势识别的第二SRS信号；
所述接收单元，进一步用于接收所述第二SRS信号从手势发出物体反射的多个路径的第二反射信号；
所述处理单元，用于确定用于手势识别的第二SRS信号，将所述接收单元接收的所述第二反射信号与所述第二SRS信号所对应的信号模板进行相关运算，得到所述第二反射信号的相关谱特征量，并将所述相关谱特征量以及对应的用户手势标识作为所述手势发出物体输入手势的训练样本数据存储到所述存储单元中。
第二方面的上述装置，利用现有技术中上行信道探测参考信号的反射信号来进行手势训练，将手势训练得到的训练样本数据存储，从而可以用 于手势识别操作。
可选的，结合第二方面，对于所述装置，
所述发射单元进一步将所述第二反射信号的相关谱特征量以及对应的用户手势标识作为训练样本数据发送给所述网络设备。
可选的，结合第二方面，对于所述装置，
所述处理单元行相关运算，包括，
滤掉发生反射时距离所述终端设备超过设定距离的反射信号。
第二方面的上述装置，过滤超过设定距离的反射信号后，可以减少非手势所反射的无关信号的干扰，大大提高手势识别的精确度。
可选的，结合第二方面，对于所述装置，
所述存储单元事先存储SRS信号集，
所述处理单元根据所述SRS信号集确定用于手势识别的SRS信号；
或者，
所述接收单元接收网络设备发送的SRS信号集的信息；
所述处理单元根据所述接收单元接收到的SRS信号集的信息确定用于手势识别的SRS信号。
可选的，结合第二方面，对于所述装置，
所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：SRS信号的发射间隔、发射带宽、发射载频和发射功率；
或者，
所述SRS信号集包括至少一个SRS信号所对应的参数值，所述参数值属于所述参数值区间。
可选的，结合第二方面，对于所述装置，
所述SRS信号所对应的参数值区间包括下述任意一项或者多项：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于 10GHz；发射功率小于等于0dBm。
第三方面，一种生成用于手势识别的信道探测参考信号SRS的方法，其特征在于，
网络设备生成用于手势识别的SRS信号集；
所述网络设备将所述用于手势识别的SRS信号集的信息发送给终端设备；
所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于10GHz；发射功率小于等于0dBm；
或者，
所述SRS信号集包括至少一个SRS信号所对应的参数值，所述参数值属于所述参数值区间。
第四方面、一种网络设备，其特征在于，适用于移动通信系统，所述网络设备包括，发射单元及处理单元，其中，
所述处理单元，生成用于手势识别的SRS信号集，并将所述SRS信号集的信息传送给所述发射单元；
所述发射单元，用于发射所述处理单元生成的所述SRS信号集的信息给终端设备；
所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于10GHz；发射功率小于等于0dBm；
或者，
所述SRS信号集包括至少一个SRS信号所对应的参数值，所述参数值属于所述参数值区间。
附图说明
为了更清楚地说明本发明实施例的技术方案，下面将对实施例描述所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1是本发明实施例提供的SRS信号集配置流程图；
图2是本发明实施例提供的手势识别方法训练流程图；
图3是本发明实施例提供的训练样本数据结构示意图；
图4是本发明实施例提供的手势识别方法流程图；
图5是本发明实施例提供的手势识别方法识别失败后处理流程图；
图6是本发明实施例提供的终端设备结构示意图；
图7是本发明实施例提供的网络设备结构示意图。
具体实施方式
本申请实施例所公开的手势识别方法的构思是，终端设备发射上行的信道探测参考信号(Sounding Reference Signal，简称SRS信号)，从用户的手或者其他用于做手势的物体上反射回终端设备，终端设备将SRS信号的反射信号的功率谱波形,反射信号的到达时间、强度、相位等测量量中的至少一个作为特征量进行手势识别。
本申请实施例涉及的网元为：网络设备和终端设备。
在LTE系统中，接入网设备是eNB，核心网设备是MME；在UMTS系统中，接入网设备是RNC，核心网设备是SGSN；在其他无线通信系统中，也有其相应的接入网设备和核心网设备。在本申请的实施例中，上述接入网设备和核心网设备相对于用户设备设备都统称为网络设备。
本领域技术人员应当知晓，本申请实施例所述的终端设备，可以是用户终端、用户设备以及移动台等设备。
实施例1
本申请实施例1对现有技术的改进点主要在：终端设备确定并发射用于手势识别的SRS信号、接收该SRS信号针对用户手势的反射信号并进行处理；在终端设备侧设置用于存储训练样本数据的数据库。下面分别进行介绍。
在无线通信领域，SRS信号主要用于信道质量估计，从而使得基站能够针对上行链路进行频率选择性调度，包括功率控制增强和终端设备的各种初始化调制和编码方案选择、定时提前等。终端设备发送的SRS信号通过网络设备特定的广播信令指示进行配置。终端设备根据网络设备指示配置SRS信号并在指定位置发送。SRS信号所占用的时频资源，如发射间隔、发射带宽、发射载频、发射功率等均由网络设备配置。
本发明实施例对SRS信号赋予了新的手势识别的功能，为了达到通过SRS信号的反射信号识别用户手势的目的，网络设备需要为终端设备配置用于手势识别的SRS信号集。
在介绍SRS信号集的配置过程之前，先介绍本发明实施例的SRS信号的一些信息。
对于SRS信号的发射间隔，用于手势识别的SRS信号的发射间隔要比传统的用于信道质量估计的SRS信号小；但是SRS信号的发射间隔设置过小，又会导致SRS信号发射过于频繁，会对其它通信信号造成干扰，同时对于终端设备的计算能力带来巨大挑战。示意性的，本发明实施例中用于手势识别的SRS信号的发射间隔可以设置为小于10毫秒，这样的发射频率可以足够识别用户的动作手势，例如，具体数值可以为5毫秒或者10毫秒，等。
对于SRS信号的发射带宽和载频，用于手势识别的SRS信号的发射带宽和载频都要比传统的用于信道质量估计的SRS信号大。示意性的，用于手势识别的SRS信号的传输带宽设置为100MHz以上，这样终端设备判断的 距离精度可以达到30cm；用于手势识别的SRS信号的载波频率设置为大于10GHz，这样终端设备判断手势改变动作的识别度能达到1cm。
对于SRS信号的发射功率，用于手势识别的SRS信号仅仅作用于终端设备附近，无需传输到与网络设备，为了节能以及避免对正常的无线通信造成干扰，其发射功率要适当减小。例如，现有技术中用于信道质量估计的SRS信号发射功率为23dBm，本申请实施例中用于手势识别的SRS信号发射功率可以下降到0dBm。
下面结合附图1介绍SRS信号集的配置过程。
步骤101，网络设备生成用于手势识别的SRS信号集。
所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：SRS信号的发射间隔、发射带宽、发射载频和发射功率等。例如，如前所述，SRS信号集可以这样设置：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于10GHz；发射功率小于等于0dBm。相应地，终端设备可以根据该用于手势识别的SRS信号集，确定当前需要发射的SRS信号所对应的参数为：发射间隔为5毫秒,发射带宽为100MHz,发射载频为10GHz,且发射功率为0dBm；也可以确定为：发射间隔为10毫秒,发射带宽为150MHz,发射载频为12GHz,且发射功率为0dBm，等等，只要满足SRS信号集的要求即可。当然，该信号集中还可以包括一组具体的SRS信号所对应的参数值，例如，SRS信号集可以这样设置：SRS信号1，发射间隔为5毫秒,发射带宽为100MHz,发射载频为10GHz,且发射功率为0dBm；SRS信号2，发射间隔为10毫秒,发射带宽为150MHz,发射载频为12GHz,且发射功率为0dBm；等等，只要满足SRS信号集的要求即可。
步骤102，网络设备将该用于手势识别的SRS信号集的信息发送给终端设备。
步骤103，终端设备接收该SRS信号集的信息，并根据该SRS信号集的 信息确定当前用于手势识别的SRS信号。
当用户触发手势输入功能时，终端设备从所述用于手势识别的SRS信号集中选择合适的SRS信号进行发射和识别。当用于手势识别的SRS信号集规定了SRS信号的参数值区间时，终端设备生成符合上述参数值区间约束条件的SRS信号；当用于手势识别的SRS信号集规定了一组具体的SRS信号所对应的参数值时，终端设备从所述一组具体的SRS信号参数值选择其中之一。可以根据具体的应用环境和需求，选择符合条件的SRS信号。
在实际应用中，终端设备中可以固化用于手势识别的SRS信号集，在此种情形下，SRS信号集配置过程可以省略，终端设备无需网络设备生成并配置SRS信号集。当终端设备需要进行手势识别时，直接从固化的用于手势识别的SRS信号集中选取SRS信号进行发射。
在本申请实施例中，终端设备可以首先经历训练阶段，记录用户某个手势的SRS信号的反射信号的图谱特征量，并作为训练样本数据存储到数据库中。
下面结合附图2介绍训练过程。
步骤201，终端设备发射SRS信号。
对于终端设备发送的所述SRS信号，如果存在网络设备的SRS信号集配置过程，则终端设备从网络设备发送的SRS信号集中选取所述SRS信号；如果不存在网络设备的SRS信号集配置过程，则终端设备从自身固化的用于手势识别的SRS信号集中选取SRS信号。具体SRS信号的选择方式，与步骤103类似，不再赘述。
步骤202，用户距离终端设备一定的距离，打出手势。
用户可以在终端设备附近打手势，例如，30cm以内。另外，本申请各实施例中提到的手势，并不限定一定是用户的手打出的手势，也可以是用户使用其他物体打出的手势，还可以是动物打出的手势。
步骤203，终端设备接收步骤201中发射SRS信号的多个路径的反射信 号，并与步骤201中发射的SRS信号所对应的信号模板进行相关运算。
在本申请实施例中，终端设备在步骤201上行链路发送SRS信号后，需要在同一频率立刻检测所述SRS信号的反射信号，这是目前终端设备不具备的功能。现有技术终端设备只检测基站下行链路，而不检测自身或其它终端设备的上行链路。因此，为了避免SRS信号的反射信号被终端设备自身发送的信号所淹没，需要全双工技术降低或消除终端设备自身发送的SRS信号对SRS信号反射信号检测的影响。例如，终端设备可以将接收信号减去发射信号，从而消除同一时间、同一信道中发送信号对接收信号的影响。
终端设备还可以通过本地相关的方法消除发送信号对接收信号的影响，通过接收到的多个路径的SRS信号反射信号与发射的SRS信号所对应的信号模板进行相关运算获得相关谱，通过相关谱可以获得反射信号的相关谱特征量，包括每条反射径的SRS信号的反射信号的相位、功率和时延。所述发射的SRS信号所对应的信号模板可以为所述发射的SRS信号自身。功率和时延反映了反射物的大小、远近等信息，终端设备可以用来识别不同的手势，尤其是静止手势。由于多普勒效应，相位的变化反映了反射物的移动速度和移动方向，终端设备也可以用来识别不同手势，尤其是移动的手势。
步骤204，滤掉超过设定距离的SRS信号的反射信号。
由于在步骤203中终端设备将收到的多个路径SRS信号的反射信号与发射的SRS信号所对应的信号模板进行相关运算，能够计算出每个路径SRS信号的反射信号的到达时间，因此可以计算得到SRS发生发射时距离终端设备的距离，在步骤204中，终端设备可以滤掉发生发射时距离终端设备比较远的SRS信号的反射信号。上述滤掉的SRS信号的反射信号由于发生发射的距离较远，极有可能不是用户的手所反射的SRS信号，因此滤掉之后可以减少无关信号的干扰，大大提高手势识别的精确度。比如，可以滤 除30cm以外的SRS信号的反射信号。所述步骤204为可选步骤，在实际应用中可以省略。
步骤205，记录滤波后的SRS信号的反射信号的相关谱特征量，将相关谱特征量以及对应的用户手势标识作为训练样本数据存储到数据库中。
如附图3所述，所述训练样本数据301包括：相关谱特征量303，以及所述相关谱特征量对应的用户手势标识302。同一个手势的训练样本数据只有一个，但是所述手势训练样本数据中的相关谱特征量可以有多个，只要重复步骤201-206即可获得多个相关谱特征量。可选的，所述多个相关谱特征量可以进行加权平均等算法处理，相关谱特征量越多，在后续的手势识别阶段的准确性就越高。所述数据库中可以存储多个如图3所述的训练样本数据，分别对应于多个用户手势。
对于所述数据库的存储位置，可以是终端设备中的非易失性存储器，也可以是网络设备中的非易失性存储器。如果存储在网络设备中，则终端设备需要使用其支持的通信链路将所述训练样本数据传送给网络设备。
另外，无论存储训练样本数据的数据库位于终端设备还是在网络设备，所述训练样本数据可以在所有终端设备间共享，并且可以使用大数据分析的方式，进一步对训练样本数据进行处理，提高识别准确性。
在本申请实施例中，终端设备经历训练阶段后，数据库中有了各个手势的训练样本数据；或者通过共享的方式从网络设备的数据库中获得了各个手势的训练样本数据；所述训练样本数据也可以是事先固化在所述终端设备数据库中。当进入手势识别阶段时，终端设备将用户手势的相关谱特征量与数据库中训练样本数据中的相关谱特征量做对比，从而识别用户的手势。
下面结合附图4介绍识别过程。
步骤401，同训练阶段步骤201，终端设备发射SRS信号。
步骤402，同训练阶段步骤202，用户距离终端设备一定的距离，打出 手势。
步骤403，同训练阶段步骤203，终端设备接收步骤401中发射SRS信号的多个路径的反射信号，并与步骤401中发送的SRS信号所对应的信号模板进行相关运算。
步骤404，同训练阶段步骤204，滤掉超过设定距离的SRS信号的反射信号。
步骤405，记录滤波后的SRS信号的反射信号的相关谱特征量，将所测得的特征量与数据库中的训练样本数据进行匹配，识别输入手势。
实际中，可以结合反射信号到达时间，反射功率，反射信号相位变化，以及反射信号径数共同联合判定手势，同时剔除环境对手势识别的影响。
由于数据库中已经存有各个手势训练样本数据，移动终端将本次记录的相关谱特征量与数据库中各个手势训练样本数据中的相关谱特征量做对比，寻找与本次相关谱特征量最接近的训练样本数据，再判断是否是本次的用户手势类型。对于寻找最接近的训练样本数据的方法，可以采用每个特征量的差值累加值来判断，差值累加值最小的为最接近的训练样本数据；如果所述差值累加值小于设定阈值，则认为该训练样本对应的用户手势类型就是本次识别的用户手势。当然，也可以采用其他的数学算法进行判断，在此不再赘述。
如果没有寻找到本次相关谱特征量对应的用户手势类型，可能是由于周边环境干扰造成识别识别，可以重复进行识别。另外，也可以进行如附图5所示的操作：
步骤501，确定未从数据库中检索到与本次相关谱特征量对应的用户手势类型，识别失败；
步骤502，提示识别手势失败；
步骤503，确定是否已经对本次手势进行过训练；
步骤504，如果在步骤503中确定已经对本次手势进行过训练，则获取 本次用户手势标识，并执行步骤505；
步骤505，将本次相关谱特征量作为输入手势训练样本数据的新相关谱特征量加入到数据库中；
步骤506，如果在步骤503中确定未对本次手势进行过训练，可选的，提示是否对本次手势新建立训练样本数据，并执行步骤507；
步骤507，如果确认新建立训练样本数据，确定新建立的训练样本数据的用户手势标识，则将本次相关谱特征量作为本次手势的第一个相关谱特征量与用户手势标识一起存储到数据库中。
根据实施例1的上述方式，无需对终端设备进行复杂的改进或增加其他额外硬件，利用现有技术中上行信道探测参考信号的反射信号来进行手势训练和识别，将手势训练得到的训练样本数据存储，从而可以用于手势识别操作，而且在进行手势识别时也不会影响到终端设备的其他通信功能，以较低的成本实现了终端设备的手势识别功能；另外，在处理反射信号时可以过滤超过设定距离的反射信号，减少非手势所反射的无关信号的干扰，大大提高手势识别的精确度。
实施例2
实施例2与实施1相对应，是一种执行实施例1描述的手势识别方法的终端设备和网络设备。
下面结合附图6和附图7对本发明实施例2的技术方案进行描述。
如附图6所示，本申请实施例所述的终端设备601包括如下部件：处理单元602，发射单元603，接收单元604，存储单元605，输入单元606，显示单元607。
其中，所述处理单元602用于终端设备各个组成单元的控制操作以及数学运算操作；所述发射单元603用于向终端设备外界发送电磁波信号；所述接收单元604用于接收外界的电磁波信号；所述存储单元605用于存储数据、程序等文件；所述输入单元606用于接收终端设备用户的输入信 息；所述显示单元607用于向终端设备的用户显示信息。
如附图7所示，本申请实施例所述的网络设备701包括：处理单元702，发射单元703，接收单元704，存储单元705。
本领域技术人员应当理解，终端设备601、网络设备701还应当包含有完成其通信功能的其它必要部件，例如电源单元等等，但由于上述部件与本发明无关，在此不再赘述。
实施例2对现有技术的改进点、SRS信号的配置信息都与实施例1相同，在此不再赘述。
下面结合附图1、附图6和附图7介绍SRS信号集的配置过程。
网络设备701的处理单元702生成用于手势识别的SRS信号集。所述信号集的内容和配置都与实施例1步骤101相同，不再赘述。
网络设备701处理单元702将生成的用于手势识别的SRS信号集传送给发射单元703，发射单元703将所述SRS信号集信息发送给终端设备601。
终端设备601的接收单元604接收所述用于手势识别的SRS信号集信息，并传送给处理单元602，所述处理单元602根据接收单元604接收到的SRS信号集信息确定当前用于手势识别的SRS信号。
当用户触发手势输入功能时，终端设备601的处理单元602从所述用于手势识别的SRS信号集中选择合适的SRS信号进行发射和识别。具体选择方法与实施例1步骤103相同。
在实际应用中，终端设备601的存储单元605中可以固化用于手势识别的SRS信号集，在此种情形下，SRS信号集配置过程可以省略，终端设备601无需网络设备701生成并配置SRS信号集。当终端设备601需要进行手势识别时，直接从存储单元605中固化的用于手势识别的SRS信号集中选取SRS信号进行发射。
在本申请实施例2中，终端设备601可以首先经历训练阶段，记录用户某个手势的SRS信号的反射信号的图谱特征量，并作为训练样本数据存 储到数据库中。
下面结合附图2、附图6和附图7介绍训练过程。
同实施例1步骤201，终端设备601的发射单元603发射SRS信号。
对于终端设备601发送的所述SRS信号，如果存在网络设备701的SRS信号集配置过程，则终端设备601从网络设备701发送的SRS信号集中选取所述SRS信号；如果不存在网络设备701的SRS信号集配置过程，则终端设备601从自身存储单元605固化的用于手势识别的SRS信号集中选取SRS信号。
同实施例1步骤202，用户距离终端设备601一定的距离，打出手势。用户可以在终端设备601附近打手势，例如，30cm以内。
同实施例1步骤203，终端设备601的接收单元604接收其发射SRS信号的多个路径的SRS信号的反射信号，处理单元602将SRS信号的反射信号与所述发送的SRS信号所对应的信号模板进行相关运算。本申请实施例2消除终端设备自身发送的SRS信号对SRS信号的反射信号检测的影响的方式与实施例1步骤203相同，在此不再赘述。
同实施例1步骤204，终端设备601的处理单元602滤掉超过设定距离的SRS信号的反射信号。本申请实施例2滤掉信号的方式与实施例1步骤204相同，不再赘述。
同实施例1步骤205，终端设备601的处理单元602记录滤波后的SRS信号的反射信号的相关谱特征量，将相关谱特征量以及对应的用户手势标识作为训练样本数据存储到存储单元605中。所述训练样本数据的组成与实施例1步骤205相同，不再赘述。
对于所述训练样本数据的存储位置，可以是终端设备601中的存储单元605，也可以是网络设备701中的存储单元。如果存储在网络设备701中，则终端设备601的处理单元需要将所述训练样本数据传送给发射单元603,并发送给网络设备701。网络设备701的接收单元704接收上述训练样本数 据并传送给处理单元702，处理单元702再将所述训练样本数据存储在存储单元705。
无论训练样本数据存储于终端设备601的存储单元605还是在网络设备701的存储单元705，所述训练样本数据可以在所有终端设备间共享，并且可以使用大数据分析的方式，进一步对训练样本数据进行处理，提高识别准确性。
在本申请实施例2中，终端设备601经历训练阶段后，存储单元605中有了各个手势的训练样本数据；或者通过共享的方式从网络设备701的存储单元705中获得了各个手势的训练样本数据；所述训练样本数据也可以是事先固化在所述终端设备601的存储单元605中的。当进入手势识别阶段时，终端设备601的处理单元602将用户手势的相关谱特征量与存储单元605中训练样本数据中的相关谱特征量做对比，从而识别用户的手势。
下面结合附图4、附图6和附图7介绍识别过程。
同实施例1步骤401，终端设备601的发射单元发射SRS信号。
同实施例1步骤402，用户距离终端设备601一定的距离，打出手势。
同实施例1步骤403，终端设备601的接收单元604接收其发射SRS信号的多个路径的反射信号，处理单元602将反射信号与所述发送的SRS信号所对应的信号模板进行相关运算。
同实施例1步骤404，终端设备601的处理单元602滤掉超过设定距离的SRS信号的反射信号。
同实施例1步骤405，终端设备601的处理单元602记录滤波后的SRS信号的反射信号的相关谱特征量，将所测得的相关谱特征量与存储单元605中的训练样本数据进行匹配，识别输入手势。处理单元602识别输入手势的方式与实施例1步骤405相同，不再赘述。
如果没有寻找到本次相关谱特征量对应的用户手势类型，可能是由于周边环境干扰造成识别识别，可以重复进行识别。另外，也可以进行如附 图5所述的操作：
步骤501，处理单元602确定未从存储单元605中检索到与本次相关谱特征量对应的用户手势类型，识别失败；
步骤502，处理单元602控制显示单元607提示识别手势失败；
步骤503，处理单元602根据输入单元606的输入确定是否已经对本次手势进行过训练；
步骤504，如果在步骤503中确定已经对本次手势进行过训练，则处理单元602根据输入单元606的输入确定本次用户手势标识，并执行步骤505；
步骤505，处理单元602将本次相关谱特征量作为输入手势训练样本数据的新相关谱特征量加入到存储单元605中；
步骤506，如果在步骤503中确定未对本次手势进行过训练，可选的，处理单元602控制显示单元607提示是否对本次手势新建立训练样本数据，并执行步骤507；
步骤507，如果处理单元602根据输入单元606的输入确认新建立训练样本数据，则处理单元602根据输入单元606的输入确认新建立的训练样本数据的用户手势标识，将本次相关谱特征量作为本次手势的第一个相关谱特征量与用户手势标识一起存储到存储单元605中。
实施例2获得的有益技术效果与实施例1相同，不再赘述。
虽然本申请实施例以用户手势识别为例对技术方案进行了介绍，但本领域技术人员应当知晓，本申请实施例的技术方案不仅仅限于用户手势的识别，也可以用于其它装置的形状、动作识别。
在本发明各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以是两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销 售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本发明的技术方案本质上或者说对现有技术做出贡献的部分，或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本发明各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(ROM，Read-Only Memory)、随机存取存储器(RAM，Random Access Memory)、磁碟或者光盘等各种可以存储程序代码的介质。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现，且不同实施例之间，在不产生矛盾的前提下，具体的实现方式可以进行组合。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另外，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口、装置或单元的间接耦合或通信连接，也可以是电的，机械的或其它的形式连接。

Claims (18)

  1. 一种手势识别方法，其特征在于，用于传输移动通信网络信号的终端设备，所述终端设备中存储训练样本数据，所述训练样本数据包括信道探测参考信号SRS的反射信号的相关谱特征量以及对应的手势标识；
    所述方法包括，
    所述终端设备发射第一SRS信号；
    所述终端设备接收所述第一SRS信号从手势发出物体反射的多个路径的第一反射信号，并与所述第一SRS信号所对应的信号模板进行相关运算，得到所述第一反射信号的相关谱特征量；
    所述终端设备将所述第一反射信号的相关谱特征量与所述训练样本数据进行匹配，识别所述手势发出物体输入的手势。
  2. 根据权利要求1所述的方法，其特征在于，
    所述训练样本数据为事先存储在终端设备中的；
    或者，
    所述训练样本数据为所述终端设备从网络设备得到的。
  3. 根据权利要求1或2任意一项所述的方法，其特征在于，所述方法进一步包括，
    所述终端设备发射第二SRS信号；
    所述终端设备接收所述第二SRS信号从手势发出物体反射的多个路径的第二反射信号，并与所述第二SRS信号所对应的信号模板进行相关运算，得到所述第二反射信号的相关谱特征量；
    所述终端设备将所述第二反射信号的相关谱特征量以及对应的手势标识作为所述手势发出物体输入手势的训练样本数据存储到所述终端设备中。
  4. 根据权利要求3所述的方法，其特征在于，
    所述终端设备进一步将所述第二反射信号的相关谱特征量以及对应的 用户手势标识作为训练样本数据发送到网络设备中。
  5. 根据权利要求1-4任意一项所述的方法，其特征在于，
    所述相关运算包括，
    滤掉发生反射时距离所述终端设备超过设定距离的反射信号。
  6. 根据权利要求1-5任意一项所述的方法，其特征在于，所述方法进一步包括，
    所述终端设备根据事先存储在终端设备中的SRS信号集确定用于手势识别的SRS信号；
    或者，
    所述终端设备接收网络设备发送的SRS信号集的信息；
    所述终端设备根据接收到的SRS信号集的信息确定用于手势识别的SRS信号。
  7. 根据权利要求6所述的方法，其特征在于，
    所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：SRS信号的发射间隔、发射带宽、发射载频和发射功率；
    或者，
    所述SRS信号集包括至少一个SRS信号所对应的参数值，所述参数值属于所述参数值区间。
  8. 根据权利要求7所述的方法，其特征在于，
    所述SRS信号所对应的参数值区间包括下述任意一项或者多项：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于10GHz；发射功率小于等于0dBm。
  9. 一种终端设备，其特征在于，适用于移动通信系统，所述终端设备包括，存储单元、发射单元、接收单元及处理单元，其中，
    存储单元，用于存储数据库，所述数据库中存储有训练样本数据，所 述训练样本数据包括信道探测参考信号SRS的反射信号的相关谱特征量以及对应的手势标识；
    发射单元，用于发射所述处理单元确定的用于手势识别的第一SRS信号；
    接收单元，用于接收所述第一SRS信号从手势发出物体反射的多个路径的反射信号；
    处理单元，用于确定用于手势识别的第一SRS信号，将接收单元接收的所述反射信号与所述第一SRS信号所对应的信号模板进行相关运算，得到所述第一反射信号的相关谱特征量，并将所述相关谱特征量与数据库中的训练样本数据进行匹配，识别所述手势发出物体输入的手势。
  10. 根据权利要求9所述的终端设备，其特征在于，
    所述训练样本数据为事先存储在所述存储单元的；
    或者，
    所述训练样本数据为所述终端设备通过所述接收单元从网络设备得到，并存入所述所述存储单元的数据库。
  11. 根据权利要求9或10任意一项所述的终端设备，其特征在于，
    所述发射单元，进一步用于发射所述处理单元确定的用于手势识别的第二SRS信号；
    所述接收单元，进一步用于接收所述第二SRS信号从手势发出物体反射的多个路径的第二反射信号；
    所述处理单元，用于确定用于手势识别的第二SRS信号，将所述接收单元接收的所述第二反射信号与所述第二SRS信号所对应的信号模板进行相关运算，得到所述第二反射信号的相关谱特征量，并将所述相关谱特征量以及对应的用户手势标识作为所述手势发出物体输入手势的训练样本数据存储到所述存储单元中。
  12. 根据权利要求11所述的终端设备，其特征在于，
    所述发射单元进一步将所述第二反射信号的相关谱特征量以及对应的用户手势标识作为训练样本数据发送给所述网络设备。
  13. 根据权利要求9-12任意一项所述的终端设备，其特征在于，
    所述处理单元行相关运算，包括，
    滤掉发生反射时距离所述终端设备超过设定距离的反射信号。
  14. 根据权利要求9-13任意一项所述的终端设备，其特征在于，
    所述存储单元事先存储SRS信号集，
    所述处理单元根据所述SRS信号集确定用于手势识别的SRS信号；
    或者，
    所述接收单元接收网络设备发送的SRS信号集的信息；
    所述处理单元根据所述接收单元接收到的SRS信号集的信息确定用于手势识别的SRS信号。
  15. 根据权利要求14所述的终端设备，其特征在于，
    所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：SRS信号的发射间隔、发射带宽、发射载频和发射功率；
    或者，
    所述SRS信号集包括至少一个SRS信号所对应的参数值，所述参数值属于所述参数值区间。
  16. 根据权利要求15所述的终端设备，其特征在于，
    所述SRS信号所对应的参数值区间包括下述任意一项或者多项：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于10GHz；发射功率小于等于0dBm。
  17. 一种生成用于手势识别的信道探测参考信号SRS的方法，其特征在于，
    网络设备生成用于手势识别的SRS信号集；
    所述网络设备将所述用于手势识别的SRS信号集的信息发送给终端设备；
    所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于10GHz；发射功率小于等于0dBm；
    或者，
    所述SRS信号集包括至少一个SRS信号所对应的参数值，所述参数值属于所述参数值区间。
  18. 一种网络设备，其特征在于，适用于移动通信系统，所述网络设备包括，发射单元及处理单元，其中，
    所述处理单元，生成用于手势识别的SRS信号集，并将所述SRS信号集的信息传送给所述发射单元；
    所述发射单元，用于发射所述处理单元生成的所述SRS信号集的信息给终端设备；
    所述SRS信号集包括至少一个SRS信号所对应的参数值区间，所述参数值区间包括下述任意一项或者多项：发射间隔小于等于10毫秒；发射带宽大于等于100MHz；发射载频大于等于10GHz；发射功率小于等于0dBm；
    或者，
    所述SRS信号集包括至少一个SRS信号所对应的参数值，所述参数值属于所述参数值区间。
PCT/CN2016/102603 2015-10-21 2016-10-19 一种手势识别方法和装置 WO2017067465A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP16856896.2A EP3358495A4 (en) 2015-10-21 2016-10-19 Gesture recognition method and device
US15/958,380 US10732724B2 (en) 2015-10-21 2018-04-20 Gesture recognition method and apparatus

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510694748.0A CN106610716B (zh) 2015-10-21 2015-10-21 一种手势识别方法和装置
CN201510694748.0 2015-10-21

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US15/958,380 Continuation US10732724B2 (en) 2015-10-21 2018-04-20 Gesture recognition method and apparatus

Publications (1)

Publication Number Publication Date
WO2017067465A1 true WO2017067465A1 (zh) 2017-04-27

Family

ID=58556667

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/102603 WO2017067465A1 (zh) 2015-10-21 2016-10-19 一种手势识别方法和装置

Country Status (4)

Country Link
US (1) US10732724B2 (zh)
EP (1) EP3358495A4 (zh)
CN (1) CN106610716B (zh)
WO (1) WO2017067465A1 (zh)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE102018211414A1 (de) * 2018-07-10 2020-01-16 BSH Hausgeräte GmbH Verfahren zum Erkennen von Bewegungsformen
US11416077B2 (en) * 2018-07-19 2022-08-16 Infineon Technologies Ag Gesture detection system and method using a radar sensor
US11129116B2 (en) * 2019-06-21 2021-09-21 Qualcomm Incorporated System for detecting an object within a transmission path
US11914093B2 (en) * 2021-12-07 2024-02-27 Microsoft Technology Licensing, Llc RF antenna scanning for human movement classification
CN116126144B (zh) * 2023-01-16 2023-08-22 上海物骐微电子有限公司 基于pdp的手势识别方法及装置、电子设备、存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070195997A1 (en) * 1999-08-10 2007-08-23 Paul George V Tracking and gesture recognition system particularly suited to vehicular control applications
US20110156999A1 (en) * 2009-12-30 2011-06-30 Industrial Technology Research Institute Gesture recognition methods and systems
CN103529787A (zh) * 2013-10-16 2014-01-22 袁达荣 基于手势控制的遥控装置及其手势操控方法
CN104408395A (zh) * 2014-06-26 2015-03-11 青岛海信电器股份有限公司 一种手势识别方法和系统
CN104898844A (zh) * 2015-01-23 2015-09-09 瑞声光电科技(常州)有限公司 基于超声波定位的手势识别与控制装置及识别与控制方法

Family Cites Families (17)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3961964A1 (en) * 2009-11-02 2022-03-02 Nokia Solutions and Networks Oy Sounding reference signal configuration
US20120016641A1 (en) * 2010-07-13 2012-01-19 Giuseppe Raffa Efficient gesture processing
CN102053702A (zh) * 2010-10-26 2011-05-11 南京航空航天大学 动态手势控制系统与方法
CN103220101B (zh) * 2012-01-19 2018-02-16 中兴通讯股份有限公司 频谱梳信令的通知、探测参考信号的发送方法与装置
US9971414B2 (en) * 2013-04-01 2018-05-15 University Of Washington Through Its Center For Commercialization Devices, systems, and methods for detecting gestures using wireless communication signals
CN103294996B (zh) * 2013-05-09 2016-04-27 电子科技大学 一种3d手势识别方法
US20140380251A1 (en) 2013-06-19 2014-12-25 Motorola Mobility Llc Method and device for augmented handling of multiple calls with gestures
US20150024763A1 (en) * 2013-07-18 2015-01-22 Electronics & Telecommunications Research Institute Methods of resource management for device-to-device communications
US20150185858A1 (en) 2013-12-26 2015-07-02 Wes A. Nagara System and method of plane field activation for a gesture-based control system
US10613642B2 (en) * 2014-03-12 2020-04-07 Microsoft Technology Licensing, Llc Gesture parameter tuning
WO2015200710A2 (en) * 2014-06-25 2015-12-30 University Of Washington Devices, systems, and methods for detecting gestures using multiple antennas and/or reflections of signals transmitted by the detecting device
CN104615244A (zh) * 2015-01-23 2015-05-13 深圳大学 手势自动识别方法及系统
US9817109B2 (en) * 2015-02-27 2017-11-14 Texas Instruments Incorporated Gesture recognition using frequency modulated continuous wave (FMCW) radar with low angle resolution
TW201730813A (zh) * 2016-02-26 2017-09-01 國立交通大學 具有深度資訊之影像處理方法及其電腦程式產品
US9965467B2 (en) * 2016-03-28 2018-05-08 Avaya Inc. Sign language communication with communication devices
US10140776B2 (en) * 2016-06-13 2018-11-27 Microsoft Technology Licensing, Llc Altering properties of rendered objects via control points
US12022448B2 (en) * 2017-08-11 2024-06-25 Qualcomm Incorporated Carrier switching for multiple carriers using the same components of a component path

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20070195997A1 (en) * 1999-08-10 2007-08-23 Paul George V Tracking and gesture recognition system particularly suited to vehicular control applications
US20110156999A1 (en) * 2009-12-30 2011-06-30 Industrial Technology Research Institute Gesture recognition methods and systems
CN103529787A (zh) * 2013-10-16 2014-01-22 袁达荣 基于手势控制的遥控装置及其手势操控方法
CN104408395A (zh) * 2014-06-26 2015-03-11 青岛海信电器股份有限公司 一种手势识别方法和系统
CN104898844A (zh) * 2015-01-23 2015-09-09 瑞声光电科技(常州)有限公司 基于超声波定位的手势识别与控制装置及识别与控制方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP3358495A4 *

Also Published As

Publication number Publication date
US20180239438A1 (en) 2018-08-23
EP3358495A4 (en) 2018-11-07
US10732724B2 (en) 2020-08-04
EP3358495A1 (en) 2018-08-08
CN106610716B (zh) 2019-08-27
CN106610716A (zh) 2017-05-03

Similar Documents

Publication Publication Date Title
WO2017067465A1 (zh) 一种手势识别方法和装置
Wang et al. Push the limit of acoustic gesture recognition
US11442550B2 (en) Methods for gesture recognition and control
Wang et al. Channel selective activity recognition with WiFi: A deep learning approach exploring wideband information
US9910490B2 (en) System and method of cursor position control based on the vestibulo-ocular reflex
US10681711B2 (en) Method for data scheduling and power control and electronic device thereof
US20130169532A1 (en) System and Method of Moving a Cursor Based on Changes in Pupil Position
US20130169531A1 (en) System and Method of Determining Pupil Center Position
KR102349598B1 (ko) 다중 대역 무선 통신을 지원하는 전자 장치 및 이의 제어 방법
CN110554831A (zh) 操作同步方法、装置、设备及存储介质
WO2022105756A1 (zh) 定位方法、装置、终端设备、基站及位置管理服务器
CN110009004B (zh) 图像数据处理方法、计算机设备和存储介质
CN109212534B (zh) 移动终端的握持姿态检测方法、装置、设备及存储介质
CN109857245A (zh) 一种手势识别方法和终端
WO2020093278A1 (zh) 一种基于多天线的手势识别方法及装置
CN108604251A (zh) 一种多媒体文件的分享方法及终端设备
CN113534946A (zh) 一种无接触手势识别方法
EP3282680B1 (en) Blowing action-based method for operating mobile terminal and mobile terminal
CN109040588A (zh) 人脸图像的拍照方法、装置、存储介质及终端
CN114943242A (zh) 事件检测方法及装置、电子设备、存储介质
CN107479771B (zh) 一种控制方法、终端及存储介质
CN116610239A (zh) 图标处理方法、智能终端及存储介质
WO2024120444A1 (zh) 模型监督方法、装置、终端、网络侧设备及可读存储介质
WO2024017242A1 (zh) 测量间隔冲突的处理方法、装置、终端及网络侧设备
WO2024169796A1 (zh) 模型监督的方法、装置及通信设备

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: 16856896

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

WWE Wipo information: entry into national phase

Ref document number: 2016856896

Country of ref document: EP