SparkLink: A short-range wireless communication protocol with ultra-low latency and ultra-high reliability.

Dear Editor, Short-range wireless communications have been widely used in our daily life, but the pursuit of a better communication experience never stops, leading to the more stringent requirements of emerging applications.1SparkLink Alliance. (2021). Performance Evaluation of Sparlink v1.0, http://www.sparklink.org.cn.Google Scholar For example, remote control applications such as telesurgery require a delay of less than 1 ms.2She C. Sun C. Gu Z. et al.A tutorial on ultrareliable and low-latency communications in 6G: integrating domain knowledge into deep learning.Proc. IEEE. 2021; 109: 204-246Crossref Scopus (110) Google Scholar Industrial closed-loop control applications such as automatic assembly lines have a reliability requirement of at least 99.999%.3Gundall M. Schneider J. Schotten H.D. et al.5G as enabler for industrie 4.0 use cases: challenges and concepts.in: 2018 IEEE 23rd International Conference on Emerging Technologies and Factory Automation. ETFA, 2018Crossref Scopus (49) Google Scholar These stringent requirements open up a new racetrack for the development of short-range wireless communication protocols. In this letter, we report a short-range wireless communication protocol, SparkLink, with ultra-low latency and ultra-high reliability. The key technologies of SparkLink and the corresponding typical applications in different scenarios, including smart cars, intelligent manufacturing, smart terminals, and smart homes, as shown in Figure 1, are introduced. To tackle the latency challenges faced by various applications, SparkLink has chosen cyclic-prefix orthogonal frequency division multiplexing (CP-OFDM) waveform with an ultra-short frame structure and a flexible scheduling scheme of time-domain resources, and has a variable transmission delay as low as 20.833 μs. To reduce the transmission latency, the length of each radio frame tends to be short. However, the reduction of frame length is constrained by the overhead symbols in each frame, especially in the time division duplex (TDD) scheme adopted by SparkLink. To tackle this issue, SparkLink designs a frame structure with 48 frames in one superframe, where each frame shares overhead symbols in the superframe. As a result, the number of overhead symbols in each frame is substantially reduced. The length of each frame is 1/48 of the superframe length and is equal to 20.833 μs. To better illustrate the frame structure, we introduce the concept of grant (G) and terminal (T) nodes. In SparkLink, a G node acts as a central transmitter and receiver of wireless radio signals, manages and schedules the transmissions of a group of T nodes. A T node transmits and receives signals according to the scheduling of the G node. Each frame of SparkLink consists of G/T symbols (information symbols carrying data transmitted from G/T nodes to T/G nodes), an SG/ST symbol (a special G/T symbol as well as overhead symbol), and GAPs (gaps between G and T symbols). Figure 2A shows an example of the frame structure. Based on this novel frame structure, SparkLink can achieve flexible latencies by selecting different transmission periods for the information carrying packets, where the minimum latency is 20.833 μs. To meet different low-latency requirements of various applications, SparkLink exploits a flexible scheduling scheme of time-domain resources that corresponds to different sizes of packets. The minimum scheduling units for small and large packets are equal to the duration of one (20.833 μs) and 6 frames (125 μs), respectively. Based on the minimum scheduling units, the transmission period can be determined. Specifically, the transmission period of small packets is equal to the minimum scheduling unit. The transmission period of large packets is a multiple of the minimum scheduling units, which can be equal to 6, 12, …, 48 frames. Figure 2A depicts the example of the scheduling scheme for small packets. The minimum duration between two G symbols colored orange is the transmission period, which is one frame (20.833 μs). To meet the stringent reliability requirements of emerging applications, SparkLink supports ultra-reliable wireless communications by adopting error-correcting code schemes and hybrid automatic repeat-request (HARQ) schemes. SparkLink adopts the Polar code to correct errors, which has been proven to achieve the channel capability for large code lengths4Arikan E. Channel polarization: a method for constructing capacity-achieving codes for symmetric binary-input memoryless channels.IEEE Trans. Inf. Theory. 2009; 55: 3051-3073Crossref Scopus (3010) Google Scholar and has a superb performance when correcting random errors. To enhance the error-correcting performance of traditional Polar codes, SparkLink adopts cyclic redundancy check (CRC)-assisted Polar codes with successive-cancellation list (SCL) decoders, where CRC can help SCL decoders to select the most likely decoding path. Figure 2B compares the reliability performance of codes used in SparkLink and Wi-Fi 6, where low-density parity check (LDPC) codes are used in Wi-Fi 6. With the same code rate, Polar codes with CRC-assisted SCL decoding algorithms always outperform LDPC codes with log likelihood ratio-based belief propagation (LLR-BP) decoding algorithms. Although the SparkLink coding scheme can reduce error rate by a large margin, errors still occur due to time variability of channels, multi-path effect, and other unpredictable interferences. To confront this issue, HARQ is deployed in SparkLink to improve the reliability while decreasing the number of retransmissions. In traditional ARQ schemes, if transmission errors occur, the receiver abandons packets with errors and requests a retransmission from the transmitter. Note that the ARQ scheme may waste useful information in the original transmission. To reuse the information, HARQ combines the information of packets in the original transmission and the retransmission and acquires channel gains from the combination. As a result, SparkLink with HARQ has a significant reliability performance gap to the ARQ scheme. Specifically, when the number of retransmissions is 3, SparkLink with HARQ obtains a 7 dB gain compared with the ARQ scheme, and the gain increases as the number of retransmissions increases.1SparkLink Alliance. (2021). Performance Evaluation of Sparlink v1.0, http://www.sparklink.org.cn.Google Scholar With the outstanding performance in terms of latency and reliability, the SparkLink protocol can be applied in numerous applications. In the following section, we introduce the typical applications supported by SparkLink in four scenarios, namely smart cars, intelligent manufacturing, smart terminals, and smart homes. Replacing wired communications by wireless communications is an important trend in smart cars. Currently, most in-car communication protocols depend on the wiring harness, such as CAN, FlexRay, and MOST.5Ixia. (2014). Automotive Ethernet: An Overview, https://support.ixiacom.com.Google Scholar,6Buscemi A. Turcanu I. Castignani G. et al.CANMatch: a fully automated tool for can bus reverse engineering based on frame matching.IEEE Trans. Veh. Technol. 2021; 70: 12358-12373Crossref Scopus (3) Google Scholar It is noted that the wiring harness is the third heaviest component in a car and comprises 50% of the labor cost for the entire car.5Ixia. (2014). Automotive Ethernet: An Overview, https://support.ixiacom.com.Google Scholar Considering the necessity of replacement and the requirement of latency and reliability in smart cars,7Clavier A.G. “Wire” versus “wireless” communication.IEEE Microw. Mag. 2004; 5: 42-44Crossref Scopus (1) Google Scholar SparkLink aims at not only replacing wired communications but also achieving performance in terms of latency and reliability comparable to wired communications. A typical example is active noise canceling (ANC), which can be used to neutralize background noise and make drivers concentrate on driving. ANC has an end-to-end transmission delay requirement of around 20 μs, which can be facilitated by SparkLink.1SparkLink Alliance. (2021). Performance Evaluation of Sparlink v1.0, http://www.sparklink.org.cn.Google Scholar Currently, SparkLink has also demonstrated the advantage in a number of in-car applications, including battery management systems (BMSs) and in-vehicle infotainment (IVI).1SparkLink Alliance. (2021). Performance Evaluation of Sparlink v1.0, http://www.sparklink.org.cn.Google Scholar Driven by industrial transformation and upgrading, intelligent manufacturing is in urgent need that has applications with strict latency and reliability requirements and offers SparkLink a great opportunity. For example, to conduct real-time and accurate control in the industrial closed-loop system, SparkLink can meet the latency requirement within the μs level and a reliability requirement of 99.999%.8Chen K.C. Lin S.C. Hsiao J.H. et al.Wireless networked multirobot systems in smart factories.Proc. IEEE. 2021; 109: 468-494Crossref Scopus (26) Google Scholar Another example is the automated guided vehicle (AGV), which is mainly used to transport goods in warehouses. SparkLink maintains a radio link in a low-latency (less than 20 ms) and reliable (99.9%) way to prevent accidents and damages of the goods. With the help of SparkLink, not only the strict requirements of applications but also the flexible deployment requirements of machineries can be fulfilled in the field of intelligent manufacturing. Smart terminals such as phones, watches, bracelets, earphones, and tablets are gaining ubiquity in our daily life. Seamless collaboration between these terminals is critical to the user experience, which puts forward high requests to short-range wireless communications. For example, for the ultimate audio experience, the lossless audio transmission between a phone and earphones is required. SparkLink can transmit the 96 kHz × 24 bits lossless audio within 10 ms and can ensure the time synchronization between the left and the right earphones within the μs level. For the ultimate gaming experience, multi-player mobile games require real-time interactions between multiple mobile terminals. SparkLink can meet the end-to-end latency requirement of less than 100 ms and can ensure a jitter of the mobile game less than 2 ms. Besides the applications with stringent requirements, SparkLink can also serve traditional applications, such as multi-screen collaboration. In smart homes, households are usually interconnected by current short-range wireless communications such as ZigBee, Wi-Fi, and Bluetooth.9Marcos Amoroso M. Moraes R. Medeiros de Araujo G. et al.Wireless network technologies for smart homes: a technical and economic analysis.IEEE Latin Am. Trans. 2021; 19: 717-725Crossref Scopus (2) Google Scholar Households such as door locks, televisions, and lights can be controlled remotely by the homeowner. However, the emerging demands for the ultimate experience in smart homes create many applications with stringent requirements for latency and reliability. For example, 8K videos, the current highest resolution available for commercial-grade videos, will power the home theater to deliver astounding entertainment experiences. To avoid a blurry screen and getting stuck playing the 8K video, SparkLink ensures high speed transmission with reliability of 99.999% and an end-to-end transmission delay of less than 10 ms.10Huawei Technology. (2021). All-optical Quality Bearing Technology for 8K Ultra HD Videos, http://www.huawei.com.Google Scholar Besides the applications with stringent requirements, traditional applications in smart homes can still be supported by SparkLink, such as visual doorbells and automatic lighting control. This paper reports a novel short-range wireless communication protocol SparkLink with ultra-low latency and ultra-high reliability. Enabled by a flexible frame structure, SparkLink has a variable transmission delay starting as low as 20.833 μs. Moreover, by adopting Polar codes and HARQ schemes, SparkLink can support innovative applications with a reliability requirement of 99.999%. Due to the outstanding performance in terms of latency and reliability, SparkLink can support lots of typical applications in a variety of scenarios, including smart cars, intelligent manufacturing, smart terminals, and smart homes. To offer ultimate service experience, the evolution of SparkLink is still challenging. In the upcoming versions, SparkLink will include extended features, such as mesh networking and high-accuracy positioning. We expect to see more exciting progress of SparkLink in the near future.