Synchronization of a transmitter and a receiver

Abstract

The invention relates to the resetting of a receiver timer of a receiver for correlating receiver time information with a time generated by a transmitter timer of a transmitter, wherein the receiver is designed to receive radio messages transmitted to it by the transmitter. During this resetting, a first message from the transmitter is received by the receiver, wherein the reception comprises a radio time period during which the receiver's radio is active. Information is determined from which the start time or end time of the radio time period can be determined. The receiver timer is then reset according to the determined information. The method is flexible and can be used in particular for so-called "sleepy devices" or receivers that are temporarily in a sleep state.

Description

[0001] The invention relates to a method for resetting a receiver timer of a receiver for correlating receiver time information with a time generated by a transmitter timer of a transmitter, wherein the receiver is designed to receive radio messages transmitted to it by the transmitter, as well as to a receiver, a transmitter and a system for carrying out a method according to the invention.

[0002] Energy-related measurement data are often provided with time information that can be used for further processing.

[0003] Methods for time synchronization between transmitters and receivers are well known. For example, German patent application publication number DE 103 05 373 A1 describes a method for synchronizing radio systems in which a receiver measures the times between transmission blocks of a transmitter to make predictions about subsequent transmission blocks and to save energy and / or perform other tasks in the meantime.

[0004] Another example is German patent application publication number DE 10 2018 004 815 A1, which describes a method for compensating for discrepancies between the timing devices of a transmitter and a receiver by determining a carrier frequency offset and calculating a correction factor. This is done to more accurately calculate receiver wake-up times.

[0005] An example of a measuring device with a radio module for wireless transmission of measurement data is the fuse described in DE 10 2018 213 522 A1. The fuse has a modular design with a section in which the transmission function is located. This modular design enables the retrofitting of existing systems, e.g., secondary grid stations, with conventional fuses. The fuse delivers measured values, such as the RMS values ​​of the current (RMS: root mean square) and temperature values ​​with a time resolution of 1 s to a central gateway (e.g., to a data exchange device as described in DE 20 2021 000 293 U1). This allows load fluctuations in a subgrid to be monitored.For applications that require the calculation of parameters such as apparent, active, and reactive power, as well as dependent variables such as the power factor cos φ and, in general, the direction of load flow, it is necessary to measure both the prevailing voltage in a phase and the current synchronously with a high temporal resolution. These parameters are calculated from the phase position of the current and voltage, which can be described as the time offset between the two signals. To determine parameters such as cos φ with sufficient accuracy, a very small time error between the data points of the current and voltage measurements is necessary. For example, if you want to achieve a phase angle accuracy of 1° at a frequency of 50 Hz, you should achieve an accuracy of 20 ms / 360 = 55.55 µs.

[0006] Due to the technical characteristics of the fuse mentioned above and the resulting low time resolution, a more precise analysis of the processes and thus the calculation of parameters such as apparent, active and reactive power as well as dependent variables such as the power factor cos φ and the load flow direction cannot be ensured despite a sufficient sampling rate. Particularly in the mathematical processing of the current signature with externally provided measured variables, a high degree of temporal synchronicity is required, since the period duration of a 50 Hz current signal is 20 milliseconds and is therefore significantly below the resolution limit of the fuse mentioned. If, for example, the fuse transmits its measurement data via a radio interface (based on, for example,Since the sensor communicates via the Zigbee protocol and is not always active due to its self-sufficient power supply (energy harvesting) (Zigbee Sleepy End Device) to save energy, existing synchronization approaches cannot be applied. Furthermore, repeated resynchronization has a significant impact on the energy consumption of the sensor system and on data traffic in the Zigbee network.

[0007] The object of the invention is to provide a method which enables the assignment of time information wirelessly transmitted from a receiver to a transmitter to local time information and thus in particular also time synchronization and which is particularly suitable for receivers which are temporarily not supplied with energy (sleepy device).

[0008] This object is achieved by a method according to claim 1, a receiver according to claim 11, a transmitter according to claim 13 and a system according to claim 14.

[0009] To improve readability of the following explanations, an explanation of the nomenclature used is useful. In the following, reference is made to a "receiver" and a "sender". These terms refer to a core element of the inventive method, namely a message (referred to below as the "first message") that is addressed from the sender to the receiver. The names were chosen with regard to the sending of this message. A central application of the invention is sensor systems that send time-stamped sensor data to a central gateway (e.g. the security system mentioned in the introduction to the description). When using a method according to the invention in this constellation, the "receiver" mentioned below would normally be the sensor system and the "sender" the gateway.The term “receiver” can be confusing in this example because typically the majority of the data traffic between the elements runs from sensor system to gateway, i.e. from receiver to sender.

[0010] The method according to the invention relates to the resetting (in the sense of an adjustment or reset) of a receiver timer (e.g., system clock) of a receiver for the correlation (e.g., synchronization) of receiver time information with a time or times generated by a transmitter timer of a transmitter, wherein the receiver is configured to receive radio messages transmitted to it by the transmitter. "Radio messages" refers to any wirelessly transmitted messages transmitted using a wireless communication protocol (Zigbee, Bluetooth, Wi-Fi, Thread, Z-Wave, etc.).

[0011] The method according to the invention is based on the consideration that in a direct transmission of a message between a transmitter and a receiver, the delay due to the radio link can be neglected for many applications, ie is quasi instantaneous.

[0012] Therefore, the transmission of a first message from the sender to the receiver can be used to reset the receiver timer. Typically, the process of sending messages is characterized by the sequence described below. The transmission process begins at a time tStart by the sender. This typically includes a preparation phase and a radio phase beginning at time tX,Start. The receiver also receives the message in a radio phase ending at time tX,End. After receiving the message, the receiver typically performs post-processing in a post-processing phase ending at time tEnd.

[0013] Both the transmitter and the receiver therefore exhibit radio activity during a time period. According to the above considerations, it is assumed that these time periods for the transmitter and receiver coincide (with sufficient accuracy), or it is sufficient for the subsequent procedure if the beginning or end of the radio time period (hereinafter also referred to as the "radio period") coincide. Let tTx,Start denote the beginning and tTx,End the end of the radio period at the transmitter, and tRx,Start the beginning and tRx,End the end of the radio period at the receiver. For synchronization, either the relation tX,Start=tTx,Start≈tRx,Start or the relation tTx,End≈tRz,End=tX,End can be used to reset the receiver timer. For this purpose, the start time or end time of the receiver's radio time period is determined, and the receiver timer is reset according to the determined start time or end time.

[0014] In principle, there are two ways to ensure that the transmitter can correlate time information generated by the receiver with local transmitter time information using a method according to the invention.

[0015] First option: The transmitter assumes that the receiver timer is reset according to a fixed rule, e.g., the time tRx,Start is set by the receiver to a fixed time known to the transmitter (e.g., zero). The receiver determines this time. For technical reasons, the receiver timer is then not reset immediately. Instead, the time period between the times tRx,Start and tEnd is determined (hereinafter, this time period is referred to as the reception duration Δtrx = tEnd - tRx,Start) and the time tEnd is recorded. At a later time t, the receiver timer can then be set to the value tnew=Δtrx+(t−tEnd) The sender can then convert time information generated by the receiver into its own time system. To do this, the sender determines the time according to relation (1) in its own system and is set up to relate time information transmitted by the receiver to this time after the procedure has been carried out. This means that time information tReceiver sent by the receiver can then be converted accordingly. tSender=tReceiver+tTx,Start be transformed into the transmitter's own time system, or the transmitter's timer is reset according to the receiver's timer (e.g., resetting to zero for time tTx,Start). In this procedure, the determined start time or end time is used to correlate the time information transmitted by the receiver with the time generated by the transmitter's timer, assuming that the start time or end time of the radio time period recorded by the receiver corresponds to the start time or end time of the transmitter's radio time period.

[0016] Second option: A conversion according to relation (3) is not necessary if the receiver timer is synchronized to the transmitter time. Therefore, the applicant prefers this second option. In order to perform the synchronization, it is necessary for the receiver to be able to determine the time of the start of the radio section tX,Start or the time of the end of the radio section tX,Ende in the receiver's system. In the case of resetting the receiver timer according to relation (3), as illustrated above using the example of the start time of the radio section, the adjustment would then look as follows: tnew=Δtrx+(t−tEnd)+tX,Start(Sender)

[0017] For this purpose, the receiver must therefore have at least one piece of time information available that allows the determination of the local time of the transmitter at the start or end of the radio time period. One possibility would be for the transmitter to determine the local time at the start or end of the radio time period and send it in a message to the receiver, who can then determine this information from the sent message. Due to the typical architecture of a transmitter (comprising a processing unit and a radio unit), this determination would usually involve a calculation from the start time and the measured time period, e.g. tTx,Start=tStart+Δttx−(tTx,End-tTx,Start) where Δttx is referred to below as the transmission duration and symbolizes the period from the start of the transmission tStart to the end of the radio segment. Since the receiver must also perform a calculation according to relation (5), calculations are required on both the transmitter and receiver sides. To ensure the fastest possible reset of the receiver timer, it may be advisable to perform measurements instead of calculations at the transmitter. The transmitter would then, for example, measure the transmission duration Δttx and transmit the values ​​of tStart and Δttx to the receiver. The receiver can then reset the timer, for example, as follows: tnew=Δtrx+(t−tEnde)+tStart+Δttx−(tRx,Ende−tRx,Start)

[0018] At least two pieces of time information (a point in time and a duration) must then be transmitted from the sender to the receiver (e.g. instead of determining (tRx,End - tRx,Start) by the receiver, (tTx,End - tTx,Start) could be measured by the sender, transmitted to the receiver and used in relation (7), i.e. the transmission of three pieces of time information, a point in time and two durations would also be conceivable). The time information could be transmitted together in one message. Preferably, however, the start time tStart is transmitted in the first message and the duration(s) in a second message. This not only has the advantage that the information on the start time tStart is available more quickly, but sending this time information is also an indication to the receiver that the message is intended to trigger a timer reset, i.e. it serves to identify the first message as the one for the purpose of time adjustment or-synchronization of the sent message. The second option is thus characterized by the receiver receiving at least one message from the sender, wherein the at least one message from the sender includes at least one piece of time information that allows the determination of the sender's local time at the start time or end time of the radio time period during which the sender is transmitting for the purpose of transmitting the first message, and by resetting the receiver timer according to the sender's local time, wherein it is assumed that the start time or end time of the radio time period registered by the receiver corresponds to the start time or end time of the sender's radio time period.

[0019] Typical architectures are those in which the receiver or transmitter is formed by a radio unit (typically an ASIC-based hardware implementation of lower layers, e.g., the MAC layer, of the OSI communication model) and a computing unit (e.g., a microcontroller). The radio unit operates with a reduced protocol stack (e.g., up to the MAC layer), which does not allow direct determination of the radio duration or radio reception duration by the radio unit. According to one embodiment of the subject matter of the invention, the duration of the radio activity is determined using a PRS system (Peripheral Reflex System) by passing a signal indicating the activity, provided at an output of the radio unit, to a timer peripheral of the computing unit.A timer can then be used to determine the radio activity period, and a capture interrupt can be used to create timestamps at the start and end of the timer using a system clock in the processing unit, thus logging the radio activity relative to the system time. "Interrupts" are a tool available in a microcontroller to react to events without having to explicitly wait in a queue for the event, and "capture interrupts" are special "interrupts" that can be used to log the time of an event (here: the start and end of the timer).

[0020] The processing unit can also have a communication stack for sending messages and log communication stack states such as the start of a message transmission and the end of a message reception via callbacks ("callbacks" are optional mechanisms that provide insight into the communication processes). Logs can be stored as a ring buffer.

[0021] The invention also relates to a receiver with a radio unit and a processing unit, which is configured to carry out a method according to the invention. This is, for example, a Zigbee sleepy device. The invention also relates to a transmitter with a radio unit and a processing unit, and to a system comprising a transmitter and receiver according to the invention, which are configured to carry out a method according to the invention.

[0022] The invention is described in more detail below using an exemplary embodiment. The figures show: Fig. 1: a schematic representation of a sensor system designed as a “sleepy device” which is designed for wireless transmission of sensor data (corresponds to a “receiver” in the sense of the above explanations), Fig. 2: a schematic setup for recording radio activity using a PRS system, Fig. 3: Time sequences when sending and receiving a message in the course of a method according to the invention and Fig. 4: the storage of events relevant for the method according to the invention as logs in a ring buffer.

[0023] In Fig. Figure 1 shows a schematic diagram of a receiver as claimed. This is a sensor system designed to monitor an electrical circuit. On the one hand, data is collected from the power grid (typically current values), and on the other hand, the power supply is also provided via this grid. The core of this sensor system is a microcontroller µC. This receives measurement data from a sensor for recording current values. The power supply to the microcontroller µC is ensured by an AC power supply unit, which is fed from the monitored power grid. The AC power supply unit generates the supply voltage required by the microcontroller µC, e.g., 3.3V DC, through functions such as rectification, filtering, smoothing, and transformation of signals obtained from the power grid. The sensor system also includes an RF radio module, via which the microcontroller µC can wirelessly transmit measurement data.Typically, these measurement data are then time-stamped for further processing. Not shown is the power supply for the sensor and the RF radio module, which can be supplied directly by the power supply unit (NT) or via the microcontroller (µC).

[0024] Since the sensor system is powered by the monitored power grid (energy harvesting), the power supply is interrupted when no current flows. An energy storage device (e.g., capacitor-based) can be provided, which then continues to guarantee a basic power supply. In any case, it is sensible to assume a state in which little or no energy is consumed when the power supply from the grid fails. In particular, the wireless function is then deactivated. In this context, one also speaks of a "sleepy device." This is defined, for example, as a terminal device or "Zigbee end device" for the Zigbee standard.

[0025] The method according to the invention can be used in particular for the synchronization of so-called "Zigbee Sleepy End Devices." "Zigbee Sleepy End Devices" are devices with communication based on the Zigbee protocol without a permanent power supply. The power supply is based, for example, on energy harvesting, e.g., from a monitored circuit, or battery operation. Typically, such devices are controlled by a microcontroller (see [see also: Fig. 1). Synchronization can be achieved using microcontroller functions. The example assumes microcontrollers from Silabs' MGM210 series (radio chips, such as those from Texas Instruments, offer similar functionality, e.g., the TI CC2620: RFCPEISL interrupt register).

[0026] After a Zigbee sleepy end device wakes up, it can report to the Zigbee coordinator. The coordinator can then initiate a synchronization process according to the invention. A key part of the process is logging, i.e., registering or recording, the activity of the radio module communicating with Zigbee and the state of the communication stacks.

[0027] The activity of the radio module is monitored using a special component of the Silabs microcontroller, the Peripheral Reflex System (PRS). The PRS system allows peripheral modules to communicate virtually directly with each other, without requiring signal processing by the microcontroller. A transmitting module then transmits a so-called "reflex signal," which is routed by the PRS system to a receiving module, which can then perform an action based on the received signal.

[0028] The PRS system thus allows hardware events to be mapped as a source to other hardware events or interrupts (i.e., a signal or signal generation). For example, radio activity or the "packet received" radio event can be routed to a microcontroller pin and the packet duration can be measured with an oscilloscope.

[0029] At the same time, the PRS system can also be used to map the events to a timer peripheral on the microcontroller. Using the "Capture Interrupt" signal from the timer peripheral, timestamps can be created using the microcontroller's system clock; this allows radio activity to be logged or recorded relative to the system time. Furthermore, the states of the microcontroller's communication stack are logged via so-called "callbacks." Callbacks are optional mechanisms that provide insight into the communication processes.

[0030] Fig. Figure 2 shows the schematic structure. A radio event, typically a transmit TX or receive RX, is sent to a PRS channel on the microcontroller (step 1). This channel can forward the event internally to the microcontroller for processing. Specifically, a timer is started at the start of the event (step 2) and runs until the end of the event. The duration of the event is calculated by recording the start and end of the event and the timer that runs in between (step 3). This procedure can be tested by having the PRS channel send a signal to an output pin during the event (step 4). The state of the output pin can then be recorded using an oscilloscope, and the duration of the state can be measured (step 5). In normal operation, test steps 4 and 5 are then omitted.

[0031] The moment a data packet is sent over a radio link from a transmitter to a receiver, both devices are synchronized and exchange data. Synchronization at this moment occurs because the receiver's antenna reads the data from the transmitter. Data transmission via radio waves occurs at nearly the speed of light; thus, the resulting delay can be neglected over short distances in Zigbee networks. If one knows when a packet was sent and when it was received, and if the activity of the transmitting and receiving radio modules has been logged in between, the delays that occur during communication can be measured, quantified, and ultimately compensated.

[0032] An example sequence of events is shown in Fig. 3. There, a Zigbee coordinator is the transmitter and a Zigbee terminal is the receiver.

[0033] The moment the sender prepares a message for transmission, it is not sent immediately; instead, the transmission is delayed. The time commonly referred to as the "transmit delay" elapses until the message is sent. Confirmation that the message has been sent is also not synchronous, but asynchronous. The same delays occur on the receiving side.

[0034] Reasons for the delays are: The radio link is busy because another device is transmitting. The microcontroller must first process the data before it can send it. The microcontroller must decode the data after receiving it. The processors are more heavily utilized, resulting in longer processing times. The radio activity and the stack events (start and end of the communication stack activity) are logged on both the sender and receiver sides.

[0035] One assumption underlying the invention is that the radio delay is negligible. Fig. 3 this means: Radio On TX (Start) ≈ Radio On RX (Start) Radio On TX (End) = Radio On RX (End) Continuous radio on TX ≈ Continuous radio on RX = Continuous radio on

[0036] According to a preferred embodiment: the sender registers the start tStart(S) of the transmission of a first message containing the value tStart(S) the transmitter measures the transmission time Δttx (time for preparation of the radio transmission and radio transmission) The transmitter measures the radio duration tFunk or “Duration Radio on TX” using a timer the receiver measures the reception time Δtrx (time for radio reception and post-processing) the receiver registers the end point tEnd of reception the receiver acknowledges receipt of the first message to the sender (in Fig. 3 not shown) the transmitter sends a second message containing the measured reception time Δtrx and the radio duration tFunk (in Fig. 3 not shown) At time t, the receiver sets its own system time to that of the sender by tneu=tStart(S)+Δttx+Δtrx−tFunk+(t−tEnde)

[0037] Alternatively, tFunk could also be measured on the receiver side and the transmission of the value in step g) could be omitted.

[0038] It should be noted that although the formula for the calculation in step h) is based on tStart(S), for the determination of the transmitter time by the receiver it is important that the time information transmitted by the transmitter allows the determination of the start time tX,Start or the end time tX,End of the radio transmission (in the transmitter's time system).

[0039] This is the case here, because tX,Start(Sender)=tStart(S)+Δttx−tFunk and tX,End(Receiver)=tStart(S)+Δttx.

[0040] With this method, it is sufficient to log transmission activities on the sender side and reception activities on the receiver side. These logging files, or logs for short, are implemented as a ring buffer. When a new event is added, the oldest is always deleted immediately. Based on the logs, the transmission and reception delays Δt are calculated. tx and Δt rx determined. In Fig.Figure 4 shows an example of this. On the left side, the time display ("Time") and the event names ("Event ID") used in the implementation of a circular buffer are shown for the transmitter and receiver. The numbers under the "Time" heading sort the events according to their sequence. In the transmitter's event queue, the number "5" corresponds to the oldest event and the number "11" to the most recent event. In the receiver's event queue, the number "4" corresponds to the oldest event and the number "10" to the most recent event. The event names have the following meaning: Sender Initiate transfer: Transfer is initiated Pre Message Send: Preparation of transmission Radio On TX: Radio is on, TX mode (transmission mode) is switched on. Radio On Tx: Radio is on, TX mode (transmit mode) is turned off. Message sent: Confirmation or acknowledgment of the completion of a transmission Recipient Radio On Rx: Radio is on, RX mode (receive mode) is switched on. Radio On RX: Radio is on, RX mode (receive mode) is turned off. Counter Callback: Callback to log an event Pre Message Receive: Preparation of completion of reception

[0041] The actual time when a message was received by the receiver is made up of the start time in the receiver (in the example above, transmitter: time step no. 7) and the two time differences Δt tx and Δt rx This allows the current time of the end device (receiver) to be reconstructed (see above), given the delays.

[0042] A verification of the method with a network of 13 Zigbee sleepy end devices and a coordinator yielded a result of 98.5% of all data points within an error bound for the synchronization error of ±200µs.

[0043] In summary, the invention has, among other things, the following advantages and properties: The solution allows the synchronization of Zigbee sleepy end devices with an accuracy in the range of ± 200µs. The approach can be integrated into existing Zigbee applications because the implementation can be integrated into the application layer. This has the crucial advantage that the method operates independently of firmware updates of the communication module. The Zigbee stack, which handles communication, does not need to be modified. A crucial part of the technical solution is logging events and reconstructing delays throughout the entire communication process. The solution is compatible with both permanently powered devices and Zigbee sleepy end devices. This allows self-contained devices to save energy because they don't always need to be active and can switch to a passive, energy-saving standby mode, while still maintaining accurate time synchronization.

Claims

[1] A method for resetting a receiver timer of a receiver for correlating receiver time information with a time generated by a transmitter timer of a transmitter, the receiver being designed to receive radio messages transmitted to it by the transmitter, comprising the steps - receiving a first message from the transmitter by the receiver, wherein the receiving comprises a radio time period during which the radio of the receiver is active, - Determining information from which the start time or the end time of the radio time period can be determined, and - Reset the receiver timer according to the information obtained. [2] Method according to claim 1, characterized by - sending the first message to the receiver by the transmitter, wherein the sending includes a radio time period during which the transmitter transmits, - Determining information from which the start time or the end time of the radio time period can be determined, and - Using the determined information to correlate time information transmitted by the receiver with the time generated by the transmitter clock, assuming that the start time or end time of the radio time period determined by the receiver corresponds to the start time or end time of the radio time period of the transmitter. [3] Method according to claim 2, characterized by Resetting the transmitter timer in accordance with information from which the start time or end time of the radio time period can be determined. [4] Method according to claim 1 or 2, characterized by - receiving at least one message from the transmitter by the receiver, wherein the at least one message from the transmitter comprises at least one time information that allows the determination of the local time of the transmitter at the start time or at the end time of the radio time period during which the transmitter transmits for the purpose of transmitting the first message, and - Resetting the receiver timer according to the local time of the transmitter, assuming that the start time or end time of the radio time period registered by the receiver corresponds to the start time or end time of the radio time period of the transmitter. [5] Method according to claim 4, characterized by that the at least one message contains the local time of the sender at the start time or at the end time of the radio time period. [6] Method according to claim 4, characterized bythat the at least one message contains time information of the transmitter, which comprises the start of the transmission process for transmitting the first message and at least one time period relating to the transmission process, and the start time or the end time of the radio time period during which the transmitter transmits for the purpose of transmitting the first message can be reconstructed by the receiver from this time information of the transmitter. [7] Method according to one of claims 1 to 6, characterized by , that - the receiver or the transmitter is formed with a radio unit and a computing unit, and - the period of activity of the radio unit is determined using a PRS system. [8] Method according to claim 7, characterized by , that - a timer is used to determine the period, and - at the start and end of the timer, a capture interrupt is used to create a time stamp using a system clock of the computing unit, thus logging the radio activity relative to the system time. [9] Method according to one of claims 7 or 8, characterized by , that - the processing unit has a communication stack for sending messages, and - States of the communication stack such as the start of the transmission of a message and the end of the reception of a message are logged via callbacks. [10] Method according to one of claims 8 or 9, characterized by that logs are stored as a ring buffer. [11] Receiver with a radio unit and a computing unit, which is arranged to carry out a method according to one of claims 1 to 10. [12] Receiver according to claim 11, characterized by that it is a Zigbee Sleepy Device. [13] Transmitter with a radio unit and a computing unit, which is arranged to carry out a method according to one of claims 1 to 10. [14] System comprising a receiver according to claim 11 or 12 and a transmitter according to claim 13, which is arranged to carry out a method according to one of claims 1 to 10.

Images