Technology Evaluation Report
Total Page:16
File Type:pdf, Size:1020Kb
TECHNOLOGY RESEARCH REPORT: A Paper Study of WuR for Low Power Wireless Sensor Networks, with Observations, Tentative Conclusions and Recommended Next Steps Prepared by Ed Spence Managing Director THE MACHINE INSTRUMENTATION GROUP Revised February, 2020 Survey of the Literature for Wake Up Radio Contents PROBLEM STATEMENT ........................................................................................................................................ 2 EXECUTIVE SUMMARY ......................................................................................................................................... 4 SCOPE .................................................................................................................................................................. 5 SURVEY ................................................................................................................................................................ 5 PROTOCOLS ....................................................................................................................................................... 12 WuR RECEIVER DESIGN ..................................................................................................................................... 13 RECTIFIERS ..................................................................................................................................................... 13 ENERGY HARVESTING .................................................................................................................................... 16 ID DECODING ................................................................................................................................................. 19 CONCLUSIONS ................................................................................................................................................... 21 RECOMMENDED NEXT STEPS ............................................................................................................................ 22 REFERENCES ...................................................................................................................................................... 25 APPENDIX .......................................................................................................................................................... 27 UHF Frequency Channels in the United States ..................................................................................... 27 January 2020 The Machine Instrumentation Group 1 Survey of the Literature for Wake Up Radio PROBLEM STATEMENT The goals for this project are shared with many wireless IoT applications – maximize battery life with minimized system cost and complexity while meeting application goals for network range, data transmission rates, sensor application support, etc. From previous studies and from the literature, we assume that radio communication (transmission as well as reception) dominates the power consumption of the node. Any technique to minimize the time that the primary radio MCU is ‘on’ is the focus of much of the discussion. From an architecture point of view, there seem to be a few fundamental options to maximize battery life beyond careful circuit design and ultra-low power (ULP) component selection – 1. Operate the radio node and gateway (or hub) on a duty cycled basis, minimizing the Standby or Off time of the radio node, with added system timing complexity and/or necessity for RTC’s (additional BOM and power consumption). Note that in this scenario, node Standby time now dominates the power consumption. 2. Minimize radio transmission time (ie with a high data rate). 3. Trade off error correction coding to avoid re-transmission vs choice of frequency band to avoid collisions, while maintaining acceptable BER. 4. Apply energy harvesting techniques, if practical, to trickle charge the battery. 5. Leverage Wake-up Radio (WuR) approaches for ‘on demand’ data transmission, replacing a duty cycling protocol. 6. There is also the potential for self-wake up at the node based on a detected condition1. Early papers discuss WuR as a replacement for duty cycling MAC protocols. Ansari [10] provide power consumption and latency results for their WuR design (with node ID) compared to a simple MAC protocol with varying duty cycles. 1 Discussion of data management power implications such as whether to process data locally in the node to determine whether a condition exceeds a threshold, transmitting smaller packets such ‘health indicators’, in order to save power by reducing transmit time is beyond the scope of this report. January 2020 The Machine Instrumentation Group 2 Survey of the Literature for Wake Up Radio Figure 1. Power consumption comparisons for their Radio Triggered Wake-up with Addressing Capabilities (RTWAC) versus a Low-Power-Listening MAC protocol (Ansari 2009). Low duty cycles to reduce power consumption quickly increase latency an unacceptable amount of time for many applications. The current application is for machine condition or status monitoring. In this use case as currently defined, there is less concern for transmission latency. In addition, there is at least one use case anticipated where an ‘on-demand’ wake up feature is desirable. This opens the door to consideration of a WuR topology, perhaps with an energy harvesting component to the architecture, by assuming that time is available to accumulate enough energy to start a transmit process. Also, since large data packets are not necessarily needed for a wake-up signal (WuS), the trade off between smaller data samples and slower bit rates can be explored, reducing transmission time and hence transmission power. For the WuR approach to be of any benefit, performance may meet one or more of the following criteria: • The WuR receive has to perform reliably over the same ranges expected for the main data transmit radio. Transmitting at a significantly lower data rate and /or a lower frequency band than the main radio may help meet this criteria. • Use simpler protocols (such as OOK), which include little or no error correction, putting the onus on the WuR receiver performance parameters such as sensitivity and signal-to-noise ratio to achieve the necessary reception range. January 2020 The Machine Instrumentation Group 3 Survey of the Literature for Wake Up Radio • If a node ID addressing feature must be included to reduce overall system power consumption, the WuR should still use a much simpler data packet than a full measurement data transmission by the primary radio MCU. • Overall power consumption of the WuR circuit has to be comparable to or lower than the main radio standby power, operating on a defined duty cycle, to be considered as a replacement for duty cycled prototol. • If an on-demand wake-up and report mode is required of the application, then the power matching target of a duty cycling protocol does not necessarily apply, instead simply minimizing the WuR power consumption for the use case , frequency or condition where it is exercised may be sufficient to justify adding the additional circuitry. EXECUTIVE SUMMARY Fundamental issues regarding the use of WuR are discussed below. Degrees of freedom for WuR design and implementation include circuit technology used (ie discrete or solid state, active components), carrier frequency band, rectifier design, antenna impedance matching circuit design and integration with energy harvesting techniques. Some observations from the literature and discussions with some authors include: • Relatively recent surveys of the literature (Piyare 2017) as well as others, identify several designs that achieve 10’s of meters of communication range consuming less than 5W’s of power. • In general, many designs use lower channel frequencies (ie ISM band) to improve transmission distance for a given amount of power. An out-of-band (O-O-B) WuR channel using a different frequency band than the main radio introduces additional complexity and BOM cost. • WuR receiver designs are a key determinant of sensitivity, hence range. Design and architecture choices often trade off power consumption (passive circuits) vs lower sensitivity or longer range (active components). • Passive rectifier topologies have been described in the literature integrated with, and solely powered by, RF energy harvesting techniques to produce a WuS. These have been described as ‘zero power’ proposals. • The overwhelming majority of cited designs use On-Off Keying (OOK), a simplified form of Amplitude Shift Keying (ASK). • The designs described in the literature generally start with the assumption that high data rates (shorter transmission times) and low latency are desirable for most IoT applications. • 100-200ksps data rates appears to be a ceiling for power consumption below 2mW. • Relaxed latency requirements open the door to integrating energy harvesting circuits with the WuR, triggering a transmit sequence once sufficient energy is acquired to do so2. 2 The so called ‘zero power’ WuR. January 2020 The Machine Instrumentation Group 4 Survey of the Literature for Wake Up Radio • Large radio IC vendors are in the processing of integrating WuR functionality into the SoC3, generally based on a standards such as BLE or WIFI. In general, the highest performing designs use dedicated or specialized ICs and other active components, rather than passive, discrete designs, with associated increase in power consumption and BOM cost. SCOPE The discussion below is restricted