Improved Presence Detection

Abstract

The present invention relates to a device and method for presence detection in the vicinity of a device. The device includes an acoustic receiver and a corresponding analysis unit for analyzing received signals, and an acoustic transmitter. The transmitter and receiver are configured to operate within a predefined acoustic frequency range. The method includes the steps of receiving and analyzing acoustic signals from the environment within the frequency range, evaluating the periodicity and amount of the signals, and identifying frequency ranges and / or time windows with a low risk of interference. The transmitter is configured to transmit a presence detection signal within the identified frequency range and time window. The receiver receives reflections from the transmitted presence detection signal, which reflections enable the analysis unit to indicate a possible presence of a user.

Description

[Technical field]

[0001] The present invention relates to electronic devices that use acoustic transducers to measure the proximity or distance to an object using ultrasonic measurements, in particular the proximity or distance to a user of the electronic device.

[0002] More particularly, the present invention relates to electronic devices that detect human gestures and hand and finger movements using acoustic measurements that use acoustic transducers to measure the proximity or distance to at least one object, and the trajectory, angle or velocity of at least one object, which in most cases is the user of the device, but may also be other objects. [Background technology]

[0003] Electronic devices having at least one processing element capable of signal processing and an audio system capable of transmitting and receiving ultrasound using input / output devices (microphones, speakers, transducers, etc.) can be used for a range of different use cases, such as detecting user proximity, gestures, hand and finger movements, presence, and falls, for example, automatic sleep modes and conditional activity.

[0004] Presence detection functionality is becoming more prevalent in many electronic devices and in a wide variety of industries and use cases. Traditionally, presence detection devices have been implemented in single-purpose presence detection devices based on infrared or ultrasonic sensors with limited processing. As a recent solution, WO 2017 / 137755 or WO 2020 / 1045628 describe units that can use existing acoustic transducers in the ultrasonic range to detect the proximity of a user.

[0005] A potential problem with acoustic presence detection is interference from other sources in the vicinity of the device. In US2020367810 and WO2019122413, background noise is preferably avoided by filtering, but in this case other devices are handled by detecting those devices in the vicinity and by adjusting the transmission signal to avoid interference from those devices. Background interference can cause false positives (FP) in the detection. This can wake up the device and increase power consumption. Summary of the Invention [Problem to be solved by the invention]

[0006] Therefore, there is a need for a device and method that avoids or reduces the number of false positives in presence detection while reducing power consumption. A solution to this problem may be obtained as set forth in the accompanying claims. [Means for solving the problem]

[0007] Thus, the present invention provides a method and device in which the surroundings are monitored at a speed sufficient to provide an overview of other sources and signals from the surroundings within the operating range of the transmitter and receiver of the present invention. This is done by analyzing signals from the environment to detect frequencies or time windows in which the signals can be transmitted and detected without interference before the signals are transmitted. This reduces the power consumption of the device. [Brief description of the drawings]

[0008] The invention will now be described in more detail with reference to the accompanying drawings, which show, by way of example, the invention. [Figure 1] FIG. 2 illustrates a device and an external transmitter. [Diagram 2] 1 is a flow chart of a main aspect of the method according to the invention; [Diagram 3]2 is a flow chart according to one embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] As shown in Figure 1, the present invention relates to an electronic device 3 using at least one speaker 2 to transmit an acoustic presence detection signal, preferably in the ultrasonic range, and at least one microphone or receiver 1 to receive a signal reflected from an object, preferably a user 10, in the vicinity of the device. The device 3 comprises a signal generator 6 which transmits the generated presence detection signal to the speaker 2, for example via an interface or codec 5. Preferably, the presence detection signal is in the near ultrasonic range, for example 20kHz-30kHz, so as not to disturb the user, but it may also be in the audible range, using specific, preferably high frequencies in a known acoustic signal, such as music playback.

[0010] The processor 9 receives the signal from the microphone 1 via an interface 4 and preferably has a connection 7 from a speaker 2 or generator 6 in order to detect the time lapse from transmission to reception and possibly other characteristics affecting the reflected signal, as described in Norwegian Patent Application No. 20211333. The echo reference signal is typically implemented in the device with an echo cancellation algorithm.

[0011] Also, as described in Norwegian Patent Application No. 20211333, a speaker protection algorithm may be included in the generator 6, the interface 5, or the speaker 2. Here, the echo reference signal 7 will include the result of the algorithm. Insofar as the speaker protection algorithm may modify the output signal due to a protection algorithm (e.g. temperature protection, excursion protection) used to protect the audio output device (such as the speaker), the echo cancellation needs to get the modified signal to operate properly. The speaker protection algorithm does not modify the ultrasonic signal, but includes provisions for performing speaker protection before the ultrasonic signal is mixed into the audio signal during or after speaker protection, without being taken into account in the speaker protection algorithm, apart from the possibility of reserving headroom for the ultrasonic signal during the speaker protection algorithm. It does not matter whether the speaker protection algorithm is running in a SoC processor (such as an audio DSP) or in a higher-end smart power amplifier with an embedded processor that drives the audio output device. In either case, an external component such as an audio framework or a smart PA needs to provide the echo reference signal.

[0012] Also, because the device is transmitting a known acoustic signal, the processing in processing topology 9 needs access to an echo reference signal. If the speaker protection algorithm modifies the acoustic signal in the ultrasonic frequency range, then the echo reference signal needs to include data in the ultrasonic range. If the speaker protection algorithm ignores the ultrasonic range and does not modify the signal in the ultrasonic frequency range, then the echo reference signal may not include these frequencies.

[0013] Receiving the echo reference signal 7 allows the receive processing to analyze what speaker protection or other processing modules have done to the signal that may have altered the original acoustic signal sent from the device. If alterations have been made, the receive processing of the device can take these alterations into account during signal processing, making it easier to filter signals from other sources 8 in the environment, as described below.

[0014] As an alternative to the device shown in FIG. 1, the speaker 2 may be located in an external device or may be an independent device that transmits a signal containing some known parameters.

[0015] In FIG. 1, it is shown that other acoustic signal generators 8 are also in the vicinity of the device, generating signals that may be received directly or indirectly by the microphone and cause false detections. As mentioned above, the present invention aims to avoid any interference from these external transmitters. This is achieved by analyzing the signals from the background and other sources 8 and using a time slot or frequency range where the signal-to-noise level is maximized. The preferred signal-to-noise level may depend on the conditions and actual situation. For example, the presence detection signal may be encoded to improve the ability of the present invention to distinguish the presence detection signal from the background and other sources. Also, the signal strength of the presence detection signal may be increased relative to the background, improving the signal-to-noise level.

[0016] Also, different filtering techniques may be applied to the transmitted and detected signals to improve the detection rate, for example adaptive filtering based on feedback 7 and monitoring of the environment based on the signal received at receiver 1. Thus, the term "low" in this context with respect to background noise means that the system is able to distinguish the presence detection signal from the background, although this may vary depending on the situation. This has a signal to noise level within a selected frequency and / or range of amplitude according to the requirements of the implemented sensor and processor.

[0017] As mentioned above, the analysis according to the present invention can be configured to find time slots and frequency ranges that allow the presence detection signal to be unjammed. The analysis can be based on finding periodic events, but more advanced analysis utilizing machine learning based on past measurements can also be used.

[0018] The analysis may also be based on recognising measured acoustic signals, e.g. music played in the environment based on access to a database, predicting available time or frequency slots based on the known evolution of the recognised sounds, and transmitting presence detection signals within those time or frequency slots.

[0019] The duty cycle (DC) according to the present invention is shown in Figure 2. The process is as follows: 01:Start DC. 02: Record and analyze environmental noise. 03: If repeated periods of low noise are detected: 04: Adjust the transmission signal so that it can be sent within the low noise period, and then start transmission. · 05: Run the detection algorithm for a given time or until a presence is detected. 06: Stop duty cycle. Or, if no repeated low noise periods are detected: 07: If a low noise frequency band is detected: 08: Adapt the transmit signal to a low noise frequency and begin transmission. 05: Run the detection algorithm for a predetermined time or until a presence is detected. 06: Stop the duty cycle. Or, if the low noise frequency band is not repeatedly detected, return to step 02.

[0020] If a period or frequency band of low noise is not detected, the duty cycle may be stopped after a predetermined period of time.

[0021] In the following, we assume that interference from other devices can cause false positives (FP) in the presence detection, but not false negatives (FN) in the presence detection, since this means that the interfering signal cancels the characteristics of the transmitted signal. False positives can occur, for example, when similar devices are in close proximity, since the interfering signal may contain similar characteristics.

[0022] To detect false positives, you have the following options: 1. Do nothing. Interference collisions may occur sporadically, causing the FP to power off the sensor. It is desirable to have no interference collisions in the next duty cycle. Preferably, random variation is added to the DC interval to avoid repeated collisions in case the interfering device or source uses a regular transmission pattern. This method has some limitations, which are not discussed here. 2. Listen for interference before transmitting ultrasound. Before DC starts emitting ultrasound, it can listen for interference or environmental noise, and if noise is detected, it will either wait for the noise to stop or terminate DC and start it again after a random time. Other advantages include avoiding disturbing other devices with interference and using less power on the terminated DC.Also, this noise detector can replace currently used noise warning nodes as it responds to all kinds of noises such as water taps, cutlery, kitchen appliances, etc. If it waits too long to reach silence, it will report a noisy environment and fall back to safe mode. 3. Listen for interference and sync up. Before starting to emit ultrasound at DC, listen for interference. If such noise is detected, check if there is a noise-free time slot between the chirps from the interfering device, and if so, synchronize the chirp transmission to fit into that slot. In this way, the probability of FP on both devices can be reduced. 4. Listen for interference after presence detection. Assuming that interference only causes FP and not FN, we can stop transmitting ultrasound after presence detection and check for interference. If we find that interference caused presence detection, we start a new DC immediately, hoping that there will be no collision next time. The only practical difference from 1 above is that the next DC will start sooner. We can randomize the wait time between trying to play ultrasound again after the interfering noise has disappeared, to avoid many devices starting at the same time just after one device has finished. 5. Combine methods 2, 3, and 4. According to one embodiment, all three methods can be combined in the processing topology on the same node.

[0023] The above options are illustrated in Fig. 3, which shows one possible embodiment of the method and procedure according to the invention, where a more detailed example of a duty cycle is shown. The duty cycle is automatically started after a predefined time schedule, for example based on the method described in Norwegian patent application no. 20210781.

[0024] In FIG. 3, the following preferred sequence is shown using reference numbers: 11.Start the duty cycle. 12. Start audio I / O in mute mode. 13, 14. Wait to estimate latency until the timestamp latency has not been determined. 15. Listen for interference or other noises. 16. Noise is detected. 17.Are there any noise-free time slots? 18.Time slot not found. 19. Has the silence timeout been reached? 20. Silence timeout not reached, return to listen mode. 21. Unmute the output. 22. Synchronize to a noise-free timeslot. 23. Set a longer time until the next duty cycle. 24. Report no timeslot found and end duty cycle. 25. Perform presence detection. 26. Discovery timeout reached? 27. If detection timeout is reached, report non-presence and end duty cycle. 28. Presence detected, mute output. 29. Listen for interference and other noises. 30. If noise is detected, set the time until the next duty cycle to be shorter. 31. If no noise is detected, the time until the next duty cycle is set randomly. 32. Report presence and end duty cycle. 33. Report presence and end duty cycle.

[0025] In a first step 12, the audio I / O is muted and the system is put into listen mode, and preferably a silence timeout period is set for unmuting the audio.

[0026] The next step 13 is to pause the duty cycle for a predetermined time until a predetermined latency period is reached (14), thereby reducing the chance of a collision and therefore increasing the latency of vacant seat detection.

[0027] After the latency period has expired, the system listens for interference or other noise (15). If noise is detected (16), it is analyzed to find noise-free time slots (17). If no noise is detected, the outgoing transmitter is unmuted (21) and presence detection begins.

[0028] If a time slot is found (17), typically if the noise is found to be periodic, the system is synchronized (22) to a low noise time slot and the output transmitter is unmuted (21) to perform presence detection. Alternatively, in step 17, the system may find a slot in a frequency range where the noise is sufficiently low, for example reducing the chirp frequency range to focus on the low noise level portion. A combination of time and frequency slots is also possible.

[0029] If no available low noise time slots are detected in the background noise or frequency range (18), the system registers the length of the mute period, compares it to a predefined silence timeout (19), and returns to listen mode (15) if the silence timeout has not been reached (20). If the timeout has been reached, the system sets the time for the next duty cycle (23), and reports the failure and terminates the duty cycle (24).

[0030] After unmuting the output transducer, presence detection may be performed (25). If no presence is detected, the system compares the elapsed time to a detection timeout (26) and repeats the presence detection routine (25) unless the timeout is reached. If the timeout is reached, the system reports that no presence is detected and ends the duty cycle (27).

[0031] If presence is detected, the output is muted again (28) and the system is set to listen for noise (29). If noise is detected, the time until the next duty cycle is set short (30). The system then reports the detected presence and ends the duty cycle (32).

[0032] If no noise is detected, a random time within a predetermined range is set before the next duty cycle begins (31), after which the system reports the presence and ends the duty cycle.

[0033] More specifically, during procedural steps 15-24, the system may respond in response to the detected noise characteristics as follows: A: If noise is detected in step 15, the system may decide to do nothing. This may apply if the noise profile is irregular and the transmitted signal may be distinguishable from the noise. B: If a lot of interference noise is detected, lower the chirp rate or increase the silence period to reduce the chance of collisions. C: If interference is too strong, change chirp type, chirp length, or use orthogonally coded signals. D: If significant interference noise from other devices is detected, send a message to all other devices to take one of the actions listed above. The message may be sent as a single sine wave, two alternating tones, two parallel tones, or any other code. E: If the device is able to move its output signal to another frequency band with no or less interference, it will do so. F. If the background audio noise is random and high in amplitude, the device may move its signal into the audible range to avoid other interfering devices.

[0034] In summary, the present invention relates to a method for presence detection in the vicinity of a device, and a computer-implemented program for carrying out said method. The device comprises an ultrasonic receiver and a corresponding analysis unit for analyzing received signals, and an ultrasonic transmitter. The transmitter and receiver are configured to operate within a predetermined frequency range. The transmitter and receiver are usually composed of separate transducers, preferably already implemented in the device, but may also be realized with a single transducer having both transmitter and receiver functions.

[0035] The method comprises: receiving and analyzing ultrasonic signals from the environment within a frequency range, evaluating the periodicity and amount of the signals, and identifying frequency ranges and / or time windows with a low risk of interference; Transmitting a signal within a specified frequency range and time window; - receiving by a receiver reflections from the transmitted signal, said reflections enabling an analysis unit to indicate the possible presence of a user; Includes.

[0036] If the analysis unit does not specify a clear time window or frequency range, the transmitted signals may be transmitted at random intervals, where the receiver is connected to the transmitter such that it is synchronized to the transmitted signals and is aware of the time and frequency range of each transmitted signal.

[0037] If the analysis unit detects periodicity in the signal received from the environment, the signal may be transmitted in a time window where there is little signal in the environment.

[0038] If the analysis unit detects a frequency range with low signal amplitude from the environment, the transmitted signal may be adapted to include the detected low noise frequency range.

[0039] Also, while monitoring for noise, if a low noise level is immediately detected a presence detect signal may be transmitted.

[0040] The invention also relates to an electronic device including a detector for detecting the presence of a person. The device preferably includes a transmitter and a receiver configured to transmit and receive acoustic signals in a selected ultrasonic range. The transmitter and receiver may be connected to an analysis unit for detecting the presence based on reflections from the transmitted signal. The analysis unit is also configured to receive signals from the environment within the selected range via the receiver, analyze periodicity or time sequences and frequency ranges from the signals from the environment, and based thereon recognize frequency ranges and / or time windows with low probability of interference with the transmitted signal. The transmitter is configured to transmit a presence detection signal within the time window and / or frequency range, where the signals from the environment are of low amplitude.

[0041] Preferably, the acoustic frequency range includes frequencies in the ultrasonic range, in particular the near ultrasonic range, and the presence detection signal is in the ultrasonic frequency range, preferably in the range 20kHz to 25kHz.

[0042] In response to signals detected from the environment, the analysis unit may be configured to perform at least one of the following: If the analysis unit does not recognize a frequency range or time window with a low risk of interference, the transmitter may be configured to transmit a signal in a random sequence and / or frequency range, where the receiver is synchronized with the transmitted signal so as to be able to receive and analyze reflections of the transmitted signal. If the analysis unit detects periodicity in the amplitude of the signal received from the environment, the transmission of the signal is sent in a time window in which there is a low signal from the environment. If the analysis unit detects a frequency range in which the signal amplitude from the environment is low, the transmitter may be configured to transmit a signal within a frequency range that includes low noise.

[0043] The analysis unit may also be configured to transmit a presence detection signal upon instantaneous detection of a low noise level.

Claims

1. A method for presence detection in the vicinity of a device, the device comprising an acoustic receiver and a corresponding analysis unit for analyzing received signals, and an acoustic transmitter, the acoustic transmitter and the acoustic receiver configured to operate within a predetermined acoustic frequency range, receiving and analysing acoustic signals from an environment within said acoustic frequency range, evaluating the periodicity and quantity of said acoustic signals, and identifying acoustic frequency ranges and / or time windows with a low risk of interference, the acoustic transmitter is configured to transmit a presence detection signal within the identified acoustic frequency range and time window, and the acoustic receiver receives reflections from the transmitted presence detection signal, which reflections enable the analysis unit to indicate a possible presence of a user; method.

2. 2. The method of claim 1, wherein if the analysis unit does not identify a clear time window or acoustic frequency range, the presence detection signals are transmitted at random intervals and the analysis results of the received acoustic signals are synchronized with the transmitted presence detection signals, and the time and frequency range of each of the transmitted presence detection signals is recognized.

3. The method of claim 1 , wherein when the analysis unit detects periodicity in the acoustic signals received from the environment, the presence detection signal is transmitted in a time window with few signals in the environment.

4. The method of claim 1 , wherein when the analysis unit detects an acoustic frequency range with a low signal amplitude from the environment, the transmitted presence detection signal is adapted to include the detected low-noise frequency range.

5. The method of claim 1 , wherein the analysis unit is configured to transmit a presence detection signal upon instantaneous detection of a low noise level.

6. The method of claim 1, wherein the acoustic frequency range includes frequencies in the ultrasonic range, and the presence detection signal is in the acoustic frequency range including frequencies in the ultrasonic range, preferably in the range of 20kHz to 25kHz.

7. 1. An electronic device comprising a detector for detecting the presence of a person, a transmitter and a receiver configured to transmit and receive, respectively, acoustic signals within a selected acoustic frequency range; the transmitter and the receiver are connected to an analysis unit for detecting a presence based on reflections from a transmitted presence detection signal; the analysis unit is configured to receive the acoustic signals from an environment via the receiver within the selected acoustic frequency range, analyze time sequences and the acoustic frequency ranges from the acoustic signals from the environment, and recognize acoustic frequency ranges and / or time windows with low amplitudes; the transmitter is configured to transmit a presence detection signal within the time window and / or the acoustic frequency range, and the acoustic signal from the environment is of low amplitude. Electronic devices.

8. 8. The electronic device of claim 7, wherein if the analysis unit does not recognize an acoustic frequency range or time window with a low risk of interference, the transmitter is configured to transmit a signal in a random sequence and / or acoustic frequency range, and the receiver is synchronized with the transmitted signal so as to be able to receive and analyze reflections of the transmitted signal.

9. The electronic device of claim 7 , wherein if the analysis unit detects periodicity in the acoustic signal received from the environment, the acoustic signal is transmitted in a time window in which there is a low signal from the environment.

10. 8. The electronic device of claim 7, wherein if the analysis unit detects an acoustic frequency range with a low signal amplitude from the environment, the transmitted presence detection signal is adapted to include the detected acoustic frequency range with low noise.

11. The electronic device of claim 7 , wherein the analysis unit is configured to transmit a presence detection signal upon instantaneous detection of a low noise level.

12. 8. The electronic device of claim 7, wherein the acoustic frequency range includes frequencies in the ultrasonic range, and the presence detection signal is in the acoustic frequency range including frequencies in the ultrasonic range, preferably in the range 20kHz to 25kHz.

13. A computer implemented software program configured to carry out the methods according to claims 1 to 5.

Images