Proximity sensor for a portable wireless device

Abstract

A proximity sensor for a portable wireless connection device, the sensor being arranged to determine whether a part of a user's body is in the vicinity of the portable wireless connection device. The sensor generates a time-averaged proximity, which is asserted when the device is brought into the vicinity of a part of the user's body for a given time, and can be periodically and briefly reset during a proximity period. An integration time, which can be compared to the integration time used in SAR tests, is such that the sensor can advantageously be used to reduce the radio power emitted by the portable device when it is in the vicinity of the body, the integration time can be obtained by a sigma / delta modulator configured as a rate compression unit.

Description

Technical Field

[0001] This invention relates to intelligent proximity sensors and circuitry for processing the output of the proximity sensors. More particularly, but not exclusively, this invention relates to a connected portable device, such as a mobile phone or tablet, equipped with such a proximity sensor and processor, and arranged to adapt to RF emitted from a radio interface to maintain a specific absorption rate (SAR), power density (PD), or any RF exposure within given limits. Background Technology

[0002] Capacitive proximity detectors are used in many modern portable devices, including mobile phones and tablets, to determine if the device is close to a user's body. This information is important in several ways: it is used to detect whether the phone is being actively operated by the user, whether the user is looking at a display in which case the displayed information can be adapted, and / or whether the device is switching from a low-power state to an active state. Importantly, this information is used to adapt the power level of the radio transmitter to comply with statutory SAR limits. Capacitive proximity detection is also used in touch-sensitive displays and panels.

[0003] Known capacitance sensing systems measure the capacitance of electrodes and detect the increase in capacitance when the device is placed close to a human body (e.g., on a hand, head, or knee). When there is no conductor nearby, the change in sensor capacitance is relatively small and typically corresponds to a few percent of the “background” capacitance seen by the sensor. Known capacitance detection systems may include a digital processor for subtracting drift and noise contributions and transmit digital values ​​of the network user’s capacitance in real time and / or a digital binary flag indicating proximity based on a programmable threshold.

[0004] Proximity sensors are used in portable wireless devices to reduce the power of the radio transmitter when the device is close to a user's body, such as when a mobile phone is moved to the ear to make a call or placed in a pocket. By reducing power only when the device is close, regulatory exposure limits can be complied with without unduly impairing connectivity, as the device can transmit at maximum power when not close to the body. This use is disclosed in EP 3402074 A1 and US 2015 / 237183 A1.

[0005] Several national and international standards set limits on radio energy exposure. These typically include spatial (mass, surface) and temporal averaging conditions. The ICNIRP standard (74, Health Physics 494 (1998)) specifies an averaging time of at least 6 minutes at 10 GHz, decreasing to 10 seconds at 300 GHz on a complex basis. The IEEE standard (IEEE StdC95.1-2019 (2019)) reduces the averaging time from 25 minutes at 6 GHz to 10 seconds at 300 GHz. The FCC (https: / / docs.fcc.gov / public / attachments / FCC-19-126A1.pdf) proposes an averaging time decreasing from 100 seconds below 2.9 GHz to 1 second above 95 GHz.

[0006] It is well known that limiting the power of radio transmitters in portable devices is used to keep the average SAR / PD value within a sliding time window below regulatory safety limits. In this approach, the actual transmission power decreases based on the monitored traffic, regardless of whether the device is near a user. These devices do not rely on proximity sensors to comply with regulatory SAR / PD limits.

[0007] Proximity sensors are typically configured to generate a timely proximity signal, i.e., a signal that is asserted immediately as soon as the sensor moves near a target object. EP 3869691 A1 discloses a sensor configured to output a time-averaged proximity signal. Summary of the Invention

[0008] The purpose of this invention is to provide a device / method that overcomes the shortcomings and limitations of the prior art.

[0009] This invention proposes a proximity sensor that, when used in portable wireless devices, can reduce the SAR dose to users of the portable devices without excessively compromising connectivity.

[0010] In contrast to known (instantaneous) methods, where RF power is reduced once the portable device is near the user, proximity sensors are configured to generate a time-averaged proximity signal that is not immediately asserted when the portable device moves near the user, but only after a certain time, if the proximity is sustained. Importantly, even when the portable device is near the user, the time-averaged proximity signal can be briefly reset to zero and then asserted again. This causes a brief increase in radio power. While this short power increase does not significantly increase the integral SAR dose, it can greatly improve data connectivity. If the proximity is only temporary, the degradation in connectivity is less, and may not occur at all.

[0011] In one variant (instantaneous / time-averaged), proximity signaling occurs once the device is brought near the user, but the power reduction is not permanent. Instead, the RF power cycles between high and low values, resulting in reduced SAR but with less impact on connectivity.

[0012] These two methods can be combined, and time-averaged proximity signals can be used to determine cyclic degradation of RF power, rather than permanent degradation. Attached Figure Description

[0013] Embodiments of the present invention are disclosed in the specification and illustrated in the accompanying drawings, wherein:

[0014] Figure 1 A capacitive proximity sensor in a portable wireless device is illustrated schematically.

[0015] Figure 2 The illustration shows the behavior of a portion of the processor of the present invention.

[0016] Figure 3 The same behavior is illustrated schematically in the form of a flowchart.

[0017] Figure 4 and Figure 5 The dose from the mobile device was plotted as a function of distance and the power as a function of time using instantaneous proximity markers and time-averaged markers.

[0018] Figure 6 and Figure 9 This is a schematic representation of a variant of the processor of the present invention.

[0019] Figure 7 and 8 The proximity signal generated by the circuit of this invention was plotted.

[0020] Figure 10 This is a schematic representation of a Sigma-Delta converter.

[0021] Figure 11 This is a schematic representation of a sigma-delta converter used as in a variant of the invention.

[0022] Figure 12 This is a schematic diagram of one embodiment.

[0023] Figure 13 This is an improved embodiment.

[0024] Figure 14 The method of the present invention is illustrated in the form of a flowchart.

[0025] Figure 15The image shows the action of a proximity sensor when a part of the body approaches a mobile device, maintains contact for a period of time, and then retracts. Detailed Implementation

[0026] Figure 1 The diagram schematically illustrates a capacitive proximity detector in a connected portable device (e.g., a mobile phone, laptop computer, or tablet computer), but the filters and methods of the present invention can be applied to different fields.

[0027] The capacitance of the detector to electrode 20 Cx Sensitive, this capacitor Cx The capacitance increases slightly when the user's hand, face, or body is close. This change caused by body proximity is masked by the capacitance of the electrode 20 itself, making it unstable. The capacitance signal is preferably amplified and processed by an analog processor 23, which may also subtract a programmable offset, and then converted into a raw digital value by an A / D converter 25. (Sample) It can be encoded as a 16-bit integer or in any other suitable format.

[0028] In non-ideal cases, the original sample It also includes noise and unwanted interference attenuated by filter 30, thus providing a range of samples useful for processing in subsequent stages. .

[0029] Preferably, the detector includes a drift correction circuit represented herein by elements 60 and 40. 160 is a baseline estimator that generates a series of samples. These samples approximate the instantaneous values ​​of the baseline, taking drift into account. This is then processed in difference unit 40 from... Subtract this value from the sample and provide the drift-corrected sample. The discriminator unit 50 then generates a binary value 'PROXSTAT', which indicates the proximity of the user's hand, face, or body. Hereinafter, the 'PROXSTAT' variable is considered a binary value. However, the invention is not limited to this and also includes detectors that generate multi-bit proximity values. The discriminator's reference input 70 is a suitable threshold, which can be predetermined at manufacturing time, defined in individual or type calibration, dynamically set by the main processor, or defined in any other way.

[0030] If the capacitive proximity sensor is part of a portable device connected for SAR control, the sensor electrode 20 will preferably be placed close to the transmitting antenna of the RF transmitter to accurately determine the distance to the wireless power source. The sensor electrode 20 can be implemented as a conductor on a printed circuit board or a flexible circuit board, and can have protective electrodes on the back and sides to suppress the detection of bodies and objects on the back or sides of the device.

[0031] In the same application, the capacitive electrode 20 can also be used as an RF antenna or a part thereof. Figure 1 This feature of the invention is illustrated. Electrode 20 is connected to decoupling capacitor C. d Connected to the radio transmitter and receiver unit 90, and having an inductor L d Alternatively, another RF blocking element can be used to block radio frequency signals. Otherwise, the radio unit 90 can be connected to an antenna separate from and independent of the sensing electrode 20; in this case, it can be directly connected to the analog interface 23 without the decoupling inductor L. d .

[0032] Figure 2 The processor is schematically shown, which processes the PROXSTAT signal 310 to determine a time-averaged TIMEAVGSTAT status flag, which is high when the user has been near the device for a period of time. TIMEAVGSTAT does not immediately rise when the user brings the phone close to their body, so a brief proximity does not cause a decrease in transmission power. The TIMEAVGSTAT status flag is raised after the phone has remained close to the user's body for a certain period of time.

[0033] In order to work, Figure 2 The circuitry incorporates some form of memory that retains a trace of the past state of the flag PROXSTAT, which is immediately close. While several variations are possible, this example features an accumulator 280 and a FIFO buffer 250. The PROXSTAT variable is available at terminal 310. Whenever a new value becomes available, the accumulator 280 adds up the values ​​of PROXSTAT and periodically resets it to zero. The time between consecutive resets is predetermined, and a granular interval is defined.

[0034] At the end of the granularity interval, before resetting accumulator 280, a new value is pushed into FIFO buffer 250 via serial input 370. If the value of accumulator 280 is zero, or below a predetermined threshold, then the value "0" is pushed into the FIFO. Otherwise, the value "1" is pushed into the FIFO.

[0035] Preferably, the length of the FIFO buffer 250 is variable and can be arbitrarily set within predefined limits. In an exemplary implementation, the buffer 250 can have a length of up to 256 positions. The granular interval between the length of the FIFO buffer 250 and each reset of the accumulator 28 defines the length of a sliding time window used to average the instantaneous proximity status flag relative to the rate at which new PROXSTAT values ​​are generated.

[0036] Note that the purpose of accumulator 280 is to slow down the insertion of new values ​​into the FIFO buffer, and thus limit the length of the FIFO buffer 250 required to obtain a given time window. The window size is determined relative to the level of integration allowed in the regulation, and accumulator 280 can be omitted if the desired window size is very short and memory is not a limiting factor.

[0037] It should also be noted that this disclosure handles a special case where the immediate status flag PROXSTAT is a one-bit value, and the contents of accumulator 280 are quantized to one bit before being pushed into the FIFO buffer. Therefore, the width of the FIFO buffer is one bit. However, this is not a necessary limitation, and the invention also includes variations where the immediate status flag PROXSTAT is a multi-bit variable, accumulator 280 accumulates a suitable function of PROXSTAT indicating whether a device is nearby, and the value pushed into the FIFO buffer 250 is also a multi-bit variable.

[0038] It should also be noted that the FIFO buffer 250 can be implemented in various ways without departing from the scope of the invention, such as using a shift register or a ring buffer.

[0039] The values ​​contained in FIFO buffer 250 are samples of the immediate state flag PROXSTAT within a sliding time window, the length of which is defined by the length of the buffer multiplied by the granular interval between consecutive introductions of new values ​​into the buffer. Adder 220 sums all values ​​in the FIFO buffer—these values ​​are single bits, the same as counting them—and compares the result in comparator 260 to a predetermined threshold 320 to produce a time-averaged proximity state flag 330. Preferably, comparator 260 has a hysteresis to avoid multiple transitions when the input value 360 ​​hovers close to the threshold 320.

[0040] While the diagram illustrates adder 220 reading all values ​​from the FIFO buffer via its respective parallel output in each loop, this is not the only way to implement a sliding summation. For example, one possible variation could include a register into which, in each loop, a new value entering the buffer at one end is added, while an old value lost at the other end is subtracted. Block 259, including FIFO buffer 25a and adder 220, can functionally be considered as an averaging or sliding summation unit. Although the illustrated variation is preferred, stable, and easy to implement, all possible implementations of an averaging or sliding summation unit can, in fact, be employed.

[0041] The time-averaged proximity state TIMEAVGSTAT can be used to modify the power of the radio transmitter in a portable device, replacing the instantaneous proximity state PROXSTAT. In a preferred variant, logic unit 270 is used to generate a combined state PROXTIMESTAT available at terminal 350, which is the result of a logical operation on PROXSTAT and TIMEAVGSTAT. The logical operation can be a logical "OR" or a logical "AND", and is preferably selected by a suitable variable PROXTIMECONFIG, corresponding to... Figure 2 The wire in the middle is 340.

[0042] Figure 3 The behavior of the invention is illustrated in a flowchart. The method begins by generating a new PROXSTAT value (step 105), which is Figure 1 The circuitry is generated at regular, periodic intervals. In step 120, accumulator 280 is updated, represented here by the variable 'TimeGranCount'. In step 122, the system checks whether the current granularity interval has been completed. In most cases, the granularity interval will not be completed, and the system will take a "No" branch, update the value of the combined state PROXTIMESTAT (step 170), and end processing until the next PROSTAT value is available.

[0043] At the end of the granularity interval, the present invention pushes a new value into the FIFO buffer (step 130), which can be "0" or "1" as disclosed above, or another suitable value. If the FIFO buffer allows multi-bit values, the sliding and TIMEAVGCOUNT are recalculated compared to the threshold TIMEAVGTHRESH (step 140), and the time averaging flag TIMEAVGSTAT is set accordingly (steps 150 and 160).

[0044] Figure 4 and 5 The illustration shows how the power of a radio transmitter can be controlled in accordance with SAR / PD limits in this invention. Figure 4 This illustrates a scenario where radio power is controlled solely by the instantaneous proximity flag PROXSTAT. The left-hand diagram shows the dose level as a function of distance for two power levels: P2 is full power, while P1 is a reduced "safe" power selected by the instantaneous proximity state PROXSTAT, which is trimmed to trigger when the distance reaches the maximum permissible dose level "L" value D1 at the nominal power. The right-hand diagram shows the power level as "P2" when PROXSTAT (line 310) is inactive, and the power level immediately drops to "P1" when PROXSTAT is active.

[0045] Figure 5The diagram illustrates a case where the output power is controlled by the combined state PROXTIMESTAT, which in this case is calculated by a logical AND operation between PROXSTAT (line 310) and TIMEAVGSTAT (line 330).

[0046] Figure 6 A variation of the invention is shown, including an AND gate 271 at the input of the averaging unit 259. The averaging unit is functionally represented as a block, and its internal structure may include... Figure 2 The FIFO buffer 250 and the adder unit 220, or have different structures. The averaging unit 259 generates a value TIMEAVGCOUNT 360, which counts the accumulated time length over a predetermined time window during which the proximity signal PROXSTAT is active. If the averaging unit... Figure 2 If implemented as disclosed, the window length will correspond to the depth of the FIFO multiplied by the update rate, which is determined by the rate at which new PROXSTAT samples are generated, and is scaled by the integral time of counter 280, if present.

[0047] In comparator 260, the value TIMEAVGCOUNT is compared with an appropriate threshold TIMEAVGTHRESH 320, as in the previous embodiment. If the threshold TIMEAVGTHRESH is exceeded and PROXSTAT is active, a time-averaged proximity flag PROXTIMESTAT 350 is generated, as indicated by logic gate 273, which in this embodiment replaces [the previous parameter]. Figure 2 270 multiplexer.

[0048] Importantly, the signal PROXTIMESTAT 350 is fed back to the input of the averaging unit via AND gate 271, the input of which is related to the PROXSTAT value and the complement of PROXTIMESTAT. In this embodiment, if the time-averaged proximity signal PROXTIMESTAT is already active, gate 271 prevents the accumulation of new PROXSTAT values. This is advantageous when the sensor is used to limit the radio power of a mobile device, as it allows power to return to a high level over short intervals throughout the detection period, rather than allowing high power only for a short period at the beginning as in the previous embodiments. The inventors have found that this approach to detection significantly improves connectivity when the detection period (the aforementioned window length) is quite long, i.e., spanning several minutes.

[0049] For example, if Figure 2The embodiment will generate a 2-minute high-power cycle at the beginning, followed by a low-power cycle for the remainder of the detection time until the device is removed. Due to the negative feedback disclosed above, this improved embodiment will provide a series of 2-minute high-power cycles alternating with the low-power cycle under similar parameters. In this way, connectivity can be maintained without excessively degrading SAR.

[0050] Manufacturers also have the flexibility to use shorter FIFO durations while still complying with SAR limits calculated over longer regulatory windows.

[0051] Figure 7 The values ​​of the instantaneous proximity state PROXSTAT (line 310), the corresponding values ​​of the "1" values ​​existing in the FIFO buffer TIMEAVGCOUNT (line 360), the threshold TIMEAVGTHRESH (line 320), and the time-averaged proximity state TIMEAVGSTAT (line 350) are shown. Digital signals 310 and 350 have been shifted by arbitrary amounts to improve readability.

[0052] Figure 7 This corresponds to a scenario where the mobile device temporarily moves closer to the user within four short intervals, as shown in the instantaneous proximity state (line 310). As explained above, this causes the TIMEAVGCOUNT value to rise, but does not reach the threshold level (line 320). Therefore, the time-averaged proximity state (line 350) remains inactive.

[0053] Figure 8 This corresponds to the situation where the proximity between the mobile device and the user is extended, and the proximity sensor is configured as follows: Figure 6 As shown. The accumulated value TIMEAVGCOUNT (line 360) steadily increases until it exceeds the threshold 320, at which point the time averaging approach state (line 350) becomes active. Gate 271 now prevents the accumulation of additional "1" values ​​in averaging cell 259. After a constant period, the high accumulated value 360 ​​drops below the threshold line 320, and the time averaging approach state becomes temporarily inactive, although it continues to approach. This cycle then repeats until the approach ends.

[0054] Figure 9 Another variation is shown, in which the PROXSTAT signal is selected by logic gate 271 and a second OR gate 272, the input of which receives the complementary value of PROXSTAT. When PROXSTAT is active, this variation operates as follows: Figure 6The operation is illustrated below. However, when PROXSTAT is inactive, averaging cell 259 is pre-filled with "1" values, which provides a faster response. Logic gates at 271 and 272 simulate the active PROXSTAT value even when PROXSTAT is inactive and fill the memory of the averaging cell accordingly. When the mobile device is brought close to the user's body part, the PROXTIMESTAT flag will immediately become active.

[0055]

[0056] Several improvements and modifications to the present invention are possible. On the one hand, Figure 2 The implementation uses counter 280 to reduce the depth of storage buffer 250. This is advantageous because rules require a long averaging time (approximately several minutes), and buffer 250 would take a very long time if updates were performed at full speed. Counters are a means of implementing granular intervals longer than the time it takes for two consecutive values ​​of the flag PROXSTAT to appear, but this is not the only possibility.

[0057] Unit 280 can be any device that converts a stream of PROXSTAT values ​​into an output stream with a lower rate, having a rate reduction ratio TIMEAVGGRAN. Each value in the output stream can be the sum of several PROXSTAT values, such as a count of quantized "1" values ​​as disclosed above, or the maximum value. In the critical case of a one-bit implementation, if all previous TIMEAVGGRAN input values ​​are '0', the rate compression unit 280 can be configured to produce '0', while in all other cases, it produces '1', which is equivalent to the maximum value.

[0058] Ideally, the TIMEAVGCOUNT variable 360 ​​should be a measure of the true average distance, or at least a measure of the true average proximity signal PROXSTAT within a sliding time window of the same size as the window specified in regulatory SAR measurements. In the United States, regulations require windows ranging from a few seconds to 100 seconds, depending on the frequency. Elsewhere in the world, the window is 6 minutes. In the device of this invention, the length of the integration window is determined by the depth (D or TIMEAVGDURATION) of the FIFO buffer 250 and the granularity M introduced by the counter 280, which is the number of PROXSTAT samples considered to create an entry in the FIFO. This invention is not limited to specific values ​​of D and M, but in typical implementations, D can be selected between 2 and 256 positions, while the granularity M can be a number between 1 and 16. Scan period T scan It can be effectively timed between tens and hundreds of milliseconds. A typical scan cycle is 100 milliseconds.

[0059] The total duration of the integration window is If the FIFO buffer D is large, meaning the FIFO resolution is high, the granularity M can be correspondingly coarser. This can be achieved by defining M=2. k With k=0:7, a wider range can be obtained; for example, the relative resolution of the product is not reduced (1 / 256).

[0060] Ideally, time-averaged triggering should:

[0061] 1) Calculate the total moving sum S[n] of the PROXSTAT flags (values ​​of 0 or 1) across D samples.

[0062] 2) Compare the moving sum with the threshold TH and set PROXTIMESTAT for the next sample.

[0063]

[0064]

[0065] One possible improvement is that when the PROXTIMESTAT flag is set, the power expectation is reduced, so the user is no longer exposed to high power even when devices are very close together (PROXSTAT[n]=1); therefore, feedback as follows can be introduced:

[0066]

[0067]

[0068]

[0069] A FIFO buffer of 250 is a suitable way to calculate the sliding sum. In a hardware implementation, a FIFO with D positions can be implemented as a shift register SR[0:N-1], the contents of which are shifted in SR[0] each time a new sample P[n] is inserted. In software, there are more possibilities, such as a structure like the hardware implementation or an indexed buffer, where data is not shifted, but the oldest data is overwritten by the latest sample. The index is incremented and wraps around as it reaches D.

[0070] For reasons of cost, area, and complexity, it is desirable to limit the depth D of the FIFO to a reasonable size (e.g., 256 samples). Due to the varying scan cycles T... scan Each time a new sample P[n] is generated, therefore the time average calculation mentioned above covers a duration of T. avg =T scan ×D. Typical value is T. scan =100 ms, D=256, which leads to Tavg =25.6 s, but the specification allows averaging over a maximum of 6 minutes (360 s). To fully utilize this duration, the FIFO depth could be extended, but the cost would be unacceptable. Furthermore, in some cases, the scan period may be much shorter than 100 ms, which would further increase the required FIFO size. It should be T scan =2ms, for example, a FIFO that can store 6 minutes of PROXSTAT data requires D=360s / 2ms=180,000 positions.

[0071] By processing each entry P[n] of the FIFO, the doubling of positions in the FIFO can be mitigated, allowing each P[n] to store an "aggregation" of N PROXSTAT values, where each PROXSTAT value is in the scan period T. scan Produced in the middle. Every Generate a new P[n]

[0072]

[0073] in It is a suitable aggregation function for M samples. In the examples disclosed so far, the aggregation is generated by the operation of the data compression unit 280. The monitoring duration of PROXSTAT is now... The challenge lies in defining the aggregation function f. Ideally, f would be the average of all M PROXSTAT flags, but this would require each entry in the FIFO to have log2(M) bits, which is still undesirable. In the example shown earlier, the function f takes the maximum value among the M PROXSTAT flags. Since the flags are binary values, a single high value of a PROXSTAT flag is sufficient to insert a "1" value into the FIFO. This is the most conservative approach, maximizing the radiation exposure margin at the expense of connectivity: by choosing the maximum value, the sum of the FIFO contents is 360 (see...). Figure 2 The sum of the M PROXSTAT flags is always higher than the true average of the PROXSTAT flags within the integral window of interest. On the other hand, the most flexible approach is to take the minimum of the M PROXSTAT flags, i.e., return a value of "0" if any of the M flags is low. In this approach, the sum of the FIFO buffers is always lower than the true average. The resulting process will enhance connectivity, but it is difficult to guarantee the level of regulatory radiation exposure.

[0074] In an advantageous embodiment, the rate compression unit 280 is configured to generate encoded values. This ensures that the sum of the values ​​stored in the FIFO buffer represents the true average value of PROXSTAT within the integral window of interest. This can be achieved in several ways, one possibility being when the compression unit 280 has... Figure 10 and 11 When the structure is such that it is the structure of a sigma-delta modulator.

[0075] Figure 10 This is a general representation of a first-order sigma-delta modulator with analog inputs. It receives the IN value, and differential unit 1004 subtracts the analog value fed back from the output via digital-to-analog converter 1024 from the input value. The result is processed by integrator 1005, which produces an accumulated value, followed by comparator 1009, which acts as a quantizer and produces a digital output. The modulator operates in discrete time, timed by a clock signal generated by latch 1026.

[0076] A sigma-delta modulator is typically followed by a decimation filter. In this example, the decimation filter consists of the FIFO buffer and summing unit disclosed above. They are known to have analog inputs and are indeed used to provide a digital representation of the analog input, but they can also function when the input is a digital value.

[0077] Figure 11 It shows the replacement of the sigma-delta modulator. Figure 2 The rate reduction unit of counter 280; we will show that this structure results in true time averaging. As in the previous example, the variable PROXSTAT 310 at the input is combined in AND gate 1010 with the complement of either the PROXTIMESTAT variable or the TIMEAVGSTAT variable from NOT gate 1015. Therefore, PROXSTAT information is only considered in the modulator when full-power transmission is allowed. Whenever PROXTIMESTAT equals '1', the modulator is fed a '0' value, and the user is exposed to a lower value (the same effect applies to low power or high distance). Elements 1010, 1004, and 1005, indicated by thick borders, operate at full rate in each scan cycle T. scan The elements 1009, 1008, and 1015, indicated by dashed borders, are updated, for example, whenever a new PROXSTAT value appears, at a rate reduced by factor M or every M scan cycles. The values ​​PROXSTAT and PROXTIMESTAT can be quantized digital values, and in this case, the operation of the summing node 1004 and integrator 1005 can be achieved by repeatedly updating the digital accumulator. (Ignoring the value of TIMEAVGSTAT fed back to gate 1010) is implemented, where Acc represents the digital value appearing at the discriminator input. The following disclosure will refer to the accumulator Acc, or equivalently to the integrator 1005.

[0078] A sigma-delta converter reduces the rate of the input stream by a given ratio. The processing performed by this circuit can also be described, or perhaps more intuitively, as a circuit configured as follows:

[0079] - Calculate each (M⋅T) scan The average value of the window, not the worst-case scenario, is then passed to a FIFO, which takes the sum of all these average values.

[0080] - indicates (M⋅T) scan The average value of the window is calculated, and this average value is still represented by a single bit. Using a sigma-delta modulator, each (M⋅T) scan The decimal part of the average value in the window is transferred to the next window.

[0081] To illustrate the above, assume the granularity interval M = 16, and the PROXSTAT value sequence is in three (M ⋅ T) scan Each of the three cycles has 5, 8, and 9 high values ​​respectively. The first entry in the FIFO will be 0 (from the maximum of 16, there are only 5 high flags), and the accumulated value of 5 is carried over to the next window. The second window sets 8 values, and the accumulated value is now 5+8=13. The second value in the FIFO is then "1", and the value "1" will be subtracted from the accumulated value of the next M=16 scan cycles. During this time, there will be 9 new high PROXSTAT flattening, so in the third (M⋅T) scan When the window ends, the accumulated value will be 13 – 16 + 9 = 6, etc.

[0082] For example, if the granularity interval is again set to M = 16 in the first loop (the 16 values ​​of PROXSTAT will be recovered from one value pushed into the FIFO buffer), then the 16 PROXSTAT values ​​will be summed into the numerical accumulator 1005. Unit 1002 is an M-factor rate compressor that, at the end of the granularity loop, forwards the accumulated sum to comparator 1009. If the accumulated sum exceeds a specified threshold, which can be 8 = M / 2, then comparator 1009 will produce an output value of "1". Unlike the accumulator in the previous example, accumulator 1005 is not reset to zero at the end of the granularity interval; instead, the output value of the sigma-delta converter is subtracted from the input and multiplied by the factor M introduced by the M-rate expander 1006. Simulations show that the average proximity value produced in this way is more reliable than the average proximity value in the previous example.

[0083] Like in Figure 2 middle, Figure 12A sigma-delta modulator with an added FIFO buffer 250 and adder 220 is shown. The dashed box 520 encloses the elements operating at a reduced rate, per (M⋅T) scan The cycle is repeated once, while the elements 1010, 1005, and 1004, indicated by thick borders, operate at full speed.

[0084] Since the threshold applied to comparator 1009 is M / 2, the minimum accumulated value occurs when the comparator outputs '1' and M subsequent PROXSTAT values ​​are '0'. The accumulator decrements M times, and after receiving these M PROXSTAT=0 values, its value is Acc. min = M / 2 - M = -M / 2. Similarly, if the accumulated value is just below M / 2, Q at latch 1008 is 0, and if the M following PROXSTAT values ​​are '1', then the highest accumulated value occurs: Acc. max =M / 2+M=3M / 2.

[0085] For simplicity, the threshold applied to trigger 1009 can be set to any value, and it can be zero. Choosing M / 2 makes the output match the true average from the first loop. Setting it to 0 will only change the behavior after a reset. The accumulator range is [-M, M].

[0086] When the FIFO buffer 250 is full, the total number of PROXSTAT values ​​applied to the modulator input is approximated by the sum of the '1' values ​​in the FIFO, 365 (S), making It is an approximation based on the FIFO content, where, This indicates the exact number of '1' values ​​at the modulator input. and The differences between them can be limited. In each (M⋅T) scan At the end of the cycle, the new value of the accumulator (Acc) is calculated. Assuming the modulator's initial state in cycle 0 is Acc = Acc[0] and Q = Q[0], then the value of the accumulator at the end of cycle 1 is:

[0087]

[0088] It can be extended until the end of the Dth (M⋅Tscan) cycle.

[0089]

[0090] Summing the terms of Acc[1:D], we get:

[0091]

[0092] Simplify, reorder, and remember.

[0093]

[0094] Generalized to any scan cycle up to index n

[0095]

[0096] is the sum of the FIFO (S) before the Q[n] shift; thus,

[0097]

[0098] In all cases, to meet the regulatory SAR limit, the highest possible value of N must be considered only based on its approximation for consideration. The minimum and maximum possible values of and ; thus, the blind scheme will require the FIFO sum S to be increased by before comparison to ensure that and the rules are always met. Alternatively, the threshold TH can be reduced by 2.

[0099] Since Acc[n] is known and stored in the current accumulator, the scheme can be improved. This value is compared with M / 2 to decide whether to increase by only 1 (when Acc[n] < M / 2), otherwise by 2. This condition is already known at the end of cycle [n] and is included in Q[n]. The upper limit of S ( ) is:[[]]

[0100]

[0101] Finally, the threshold for accumulator quantization can be set to any value. The easiest to implement is zero. The reason for initially choosing M / 2 is that it matches the behavior of the exact / match case and corresponds to the definition of the average considering the output signal to be 0 / 1 (i.e., having a bias of 0.5). Changing from M / 2 to 0 makes the first TIMEAVGSTAT decision (when the FIFO is first filled) pessimistic by ½, and this term will disappear in subsequent decisions.

[0102] Figure 13 Shows an improved version of the processor, where the accumulator threshold in the modulator is set to zero, and the FIFO sum is limited below the maximum value. As mentioned before, the dashed box 520 represents the elements timed at a reduced rate.

[0103] This method can be performed by Figure 14The flowchart is used to represent this. Step 402 is an initialization process performed only at the start of processing, and among other things, it may specifically include initializing the FIFO, resetting the output Q of latch 1008 to 0, resetting the accumulator (the output of integrator 1005 after rate downconversion 1002), and setting various counters to zero. The successive steps will refer to counter m, which repeatedly increments from 0 to M and is used to determine (M ⋅ T). scan The position within the cycle. This counter can also be reset to zero during initialization 302.

[0104] In step 405, the method waits until a new PROXSTAT value is available. In principle, the method waits for each... T scan It is executed during the interval and loops back to step 305.

[0105] Step 420 represents the update of the accumulator value Acc, or equivalently, the action of integrator 1005 and adder node 1004.

[0106]

[0107] Where ∧ represents the AND operation. In step 422, the counter m mentioned earlier is incremented.

[0108] Test 455 is used to determine (M⋅T) scan The cycle ends. The accumulator is updated M times until control is passed to test 440, which represents the action of comparator 1009 and latch 1008. After that, the sum of FIFO is calculated (step 460), the new value of Q is pushed into FIFO (step 470), comparator 260 compares the sum value with the threshold (step 475) and sets the TIMEAVGSTAT value accordingly (steps 480, 481), the counter m is reset to zero, the value of PROXTIMESTAT is updated (step 490), and the cycle starts again.

[0109] Figure 15 The operation of the proximity sensor of the present invention in a smartphone is illustrated. The device is tested through the following steps: a simulated body part is brought close to the smartphone, and the electromagnetic radiation absorbed in the simulated body part is measured. During the test, the distance between the simulated body part and the smartphone is also measured. dInitially, the distance decreases from 50 mm to 0 mm (area A in the diagram), then the phantom maintains contact with the phone for a period of time (area B), and finally moves away until the distance is again 50 mm (area C). The dashed line "PS Off" indicates the dose when the phone has no mitigation mechanism: the radio power is not adapted at all. The dashed line "TA Off" is the response obtained when the RF power is controlled by an instant proximity indicator like PROXSTAT. When the distance between the approaching phantom and the phone reaches a certain threshold, the power suddenly decreases, and when the phantom is at the same distance in its backward movement, the power returns to its initial value.

[0110] The solid line TA-PS is the result of the time-averaged proximity flag of this invention. A short time after the approaching phantom crosses the threshold, the RF power decreases, but it briefly rises back to full value in a cyclical manner, even during phantom contact. In this way, the time-integrated dose is reduced, but connectivity is not degraded as in the previous case.

[0111] In another embodiment not shown in the figures, the proximity detector of the present invention can omit the average FIFO and generate a cyclic proximity signal that causes the RF power to cyclically decrease as the portable device approaches the user. In this way, the RF power alternates between high and low values ​​according to a simple periodic rule with a defined duty cycle. This reduces SAR with minimal impact on connectivity. When an immediate proximity indicator indicates that the wireless device is near the user, the radio power is periodically set to a lower value and then returns to a higher normal value. Simplicity is a significant advantage of this variant, making it ideal for low-cost devices.

[0112] In the example above, the variable applied to input 310 is the instantaneous proximity state PROXSTAT, which is generated once the proximity sensor senses that a given capacitive electrode is near a given RF antenna. However, the proximity sensor may actually be able to determine the proximity relative to several antennas and capacitive input electrodes of the portable device. In an advantageous variant, the proximity sensor is configured to establish a logical combination of proximity signals from several antennas and / or input electrodes, and this combined signal is presented at input 310 for time averaging. For example, the signal could be the sum of proximity signals from multiple antennas / electrodes, a logical "OR", or any suitable logic function.

[0113] Proximity sensors can be configured to distinguish between objects with high dielectric constants and objects with low dielectric constants, the former indicating the presence of a part of a user's body. Other proximity sensors have means of distinguishing between legitimate proximity and contaminants, such as dew or water on a portable device. In proximity sensors with these capabilities, the signal provided to the input of the time averaging unit 310 can be a qualified proximity signal BODYSTAT, which is generated in response to proximity to a body-like object and otherwise not generated. Several devices can establish a qualified proximity signal that is generated when a part of a user's body is nearby, but are fairly or substantially insensitive to inanimate bodies or contaminants, and all of these are included in this invention. Such qualified proximity signals most commonly take the form of binary values, but they can also be multi-level.

[0114] As described above, a key implementation of the invention operates on a one-bit binary signal, but this is not the only case. In fact, the invention can be applied to operate on non-binary signals, as long as they can be represented by a digital word of an appropriate number of bits. For example, a proximity sensor can be configured to generate multi-level instantaneous proximity signals, where different values ​​correspond to multiple distances. In this case, the distance levels can be encoded in binary form, for example, with two-bit words capable of representing four distance levels from 0 (not close) to 3 (closest), which are presented at input 310. Naturally, the rate compressor 280, the FIFO buffer 250, and the summing function 220 will have suitable bit widths.

[0115] In another variation of the invention, the value presented at input 310 for time averaging may be a digital value D(n) representing the self-capacitance of the antenna / electrode, rather than the binary value PROXSTAT, which is determined by the ADC and is after filtering and baseline subtraction.

[0116] This invention uses a storage buffer to implement window averaging, which has as many taps (M or TIMEAVGRAN) as a scaling factor as the duration of the averaging interval is required. While accurate, this requires a considerable amount of memory. The averaging unit can comprise a general-purpose digital low-pass filter, such as a recursive low-pass digital filter (IIR filter), which can be implemented more compactly, instead of a sliding window averager with a FIFO buffer.

[0117] This disclosure also includes appendices “Time Averaging 2.0” and “Time Averaging Proximity Sensor”.

[0118] Reference symbols in the attached figure

[0119] Ld decoupling inductor

[0120] Cx tactile capacitance

[0121] Cd decoupling capacitor

[0122] 20 electrodes

[0123] 23. Simulated Processor

[0124] 25 Analog-to-Digital Converters

[0125] 30 Filters

[0126] 40 differences

[0127] 50 Discriminators

[0128] 60 Baseline Estimator

[0129] 70 threshold

[0130] 90 RF receiver / transmitter

[0131] 105 Generation of new PROXSTAT values

[0132] 120 count

[0133] 122 Particle size interval ends

[0134] 130 Push into shift register

[0135] 135 Update TIMEAVGCOUNT

[0136] Comparison of 140 and TIMEAVGTHRESH

[0137] 150 Setting time average approach mark

[0138] 160 Reset time average close to the mark

[0139] 170 Logical Operations

[0140] Summation of 220

[0141] 230 Baseline Estimation

[0142] 240 drift correction samples

[0143] 250 FIFO buffer

[0144] 260 comparator

[0145] 270 Logical Operations

[0146] 271 Logic Gates

[0147] 272 Logic Gates

[0148] 273 Logic Gates

[0149] 280 counter or rate compression unit

[0150] 310 PROXSTAT variable

[0151] 320 TIMEAVGTHRESH variable

[0152] 330 TIMEAVGSTAT variable

[0153] 340 PROXTIMECONFIG variable

[0154] 350 PROXTIMESTAT variables

[0155] 360 TimeAvgCount variable

[0156] 365 True Average

[0157] Serial input of 370 FIFO buffer

[0158] 402 Initialization

[0159] 405 Generates a new PROXSTAT value

[0160] 420 Update accumulator / points

[0161] 422 Increment Counter

[0162] 435 Test Counter

[0163] 440 Test Accumulator

[0164] 451 Setting Q

[0165] 450 Reset Q

[0166] 460 Summation FIFO

[0167] 470 Push the new value of Q into FIFO

[0168] 475 test totals

[0169] 480 Reset TIMEAVGSTAT

[0170] 481 Setting TIMEAVGSTAT

[0171] 490 Set PROXTIMESTAT, logical operation

[0172] 520 Reduction Rate

[0173] 1002 rate downconversion

[0174] 1004 difference

[0175] 1005 points

[0176] 1006 rate upconversion

[0177] 1007 Delay

[0178] 1008 Latch

[0179] 1009 trigger

[0180] 1010 and the door

[0181] 1015 NOT gate, inverter

[0182] 1024 Digital-to-Analog Converter

[0183] 1026 Clock.

Claims

1. A proximity sensor for a portable wireless connectivity device, the proximity sensor being arranged to determine whether a part of a user's body is near the portable wireless connectivity device, the proximity sensor including processing circuitry that generates an instantaneous proximity status indicator that becomes active when a part of the user's body approaches the proximity sensor, characterized in that... An averaging unit configured to generate a true average value of an instantaneous proximity status flag within a predetermined time window, and a decision unit that generates a time-averaged proximity status flag based on the true average value, wherein the averaging unit includes a rate compression unit and a FIFO buffer that is periodically provided with values ​​provided by the rate compression unit, wherein the averaging unit includes a sigma-delta modulator configured as a rate compression unit; wherein the rate compression unit is configured such that the sum of values ​​stored in the FIFO buffer represents the true average value of the instantaneous proximity status flag within the time window.

2. The proximity sensor according to claim 1, wherein the decision unit is configured to switch the time-averaged proximity status flag to an active state when the true average value exceeds a predetermined threshold.

3. The proximity sensor according to claim 1, wherein the decision unit is configured to temporarily and repeatedly switch the time-averaged proximity status flag to an inactive state when the instantaneous proximity status flag is active.

4. The proximity sensor according to claim 1, wherein, The instantaneous proximity status flag is any of the following: a binary proximity status flag indicating that the proximity sensor is near the conductor; a qualified binary proximity status flag indicating that the proximity sensor is near a part of the user's body but has lower sensitivity to inanimate objects; a multi-level proximity status flag encoding the distance to the conductor; a digital value derived from the conversion of the self-capacitance of the sensing electrode and / or the self-capacitance of the radio antenna; or a combined proximity status flag generated by a logic function of individual proximity status flags, each of which is derived from the self-capacitance of a different sensing electrode or radio antenna.

5. The proximity sensor according to claim 1, wherein, The instantaneous proximity status flag is any of the following: a binary proximity status flag indicating that the proximity sensor is near a conductor; a qualified binary proximity status flag indicating that the proximity sensor is near a user's body but has low sensitivity to contaminants that could contaminate portable wireless connectivity devices; a multi-level proximity status flag encoding the distance to the conductor; a digital value derived from the conversion of the self-capacitance of the sensing electrode and / or the self-capacitance of the radio antenna; or a combined proximity status flag generated by a logic function of individual proximity status flags, each of which is derived from the self-capacitance of a different sensing electrode or radio antenna.

6. The proximity sensor according to claim 4, wherein the FIFO buffer has an optional length.

7. The proximity sensor of claim 1, further comprising logic circuitry configured to prevent the transmission of other values ​​of the instantaneous proximity status flag to the averaging unit if the time-averaged proximity status flag is active.

8. The proximity sensor of claim 5, in combination with a portable wireless connectivity device including a radio transmitter, wherein the proximity sensor is operatively arranged to reduce the power of the radio transmitter based on a combined proximity status flag or the instantaneous proximity status flag or the time-averaged proximity status flag value.

9. The proximity sensor of claim 1, wherein the sensor is a capacitive sensor arranged to determine whether a user is near the portable wireless connectivity device based on capacitance detected by sensing electrodes.

10. The proximity sensor according to claim 9, wherein the sensing electrode is also an antenna for transmitting radio waves.

11. A method for reducing the dose of integrated SAR for users of portable wireless connected devices, comprising, in a time loop: Obtain an instant proximity status indicator that indicates whether a portable wireless connection device is briefly near the user; The true average value of the instantaneous proximity status flag is calculated over a predetermined time window based on the current and past values ​​of the instantaneous proximity status flag, and the time-averaged proximity status flag is determined accordingly. When the time-average proximity status flag is active, reduce the power of the radio transmission of the portable wireless connectivity device; The method further includes providing the instantaneous proximity flag to a second rate lower than a first rate of the time cycle before calculating the true average value by reducing the rate of the instantaneous proximity flag to a second rate lower than a first rate of the time cycle. The calculation of the true average value involves cyclically pushing the value of the instantaneous proximity status flag into the FIFO buffer and calculating the sliding window average value based on the sum of all values ​​in the FIFO buffer.

12. The method of claim 11, further comprising briefly resetting the power to an initial value during the period in which the immediate proximity status flag is active.

Images