Integrated Sensing and Communication

Integrated Sensing and Communication (ISAC), also referred to as Joint Communication and Sensing (JCAS), is a technology candidate with promising potential. ISAC integrates sensing and spatial location of passive (not connected) objects into the mobile communication network, expanding the network's functionality beyond just communication. Reusing most of the network equipment deployed for communication can enable a smooth introduction of ISAC.  

ISAC has been the topic subject of both academic and industrial research for some years. The standardization body 3GPP has also taken the initial steps toward standardization of ISAC – notably, the 3GPP technical specification group “Service and System Aspects” released a technical report [1] and specification [2] on use cases and requirements in their Release 19.

ISAC use cases

32 potential use cases for ISAC are listed in the 3GPP technical report. Among them are use cases related to uncrewed aerial vehicles (UAVs), transportation, and industries.

ISAC for UAVs

Detection and tracking of UAVs is currently one of the most discussed applications of ISAC. A concrete example relates to the protection of critical infrastructure where the already deployed mobile communication networks are used to detect UAVs in prohibited airspace. ISAC could also be used to track “friendly” drones and provide assistance data to the UAV traffic management system for navigation and collision avoidance. In many jurisdictions, UAVs must send beacons to avoid collisions among UAVs and ISAC could complement such a system.

ISAC for transportation and automotive

ISAC can also play an important role in smart transportation. Most modern cars are equipped with onboard sensors – radar, lidar, and cameras just to mention a few – which help the car to get an overview of its immediate surroundings. However, onboard systems fail to provide information further away and cannot “see around the corner”. An ISAC system can complement onboard sensory data with more “global” data that cannot be obtained by sensors mounted on the vehicle. Another automotive-related example is the detection of pedestrians or animals on highways.

Non-automotive transportation such as rail transportation can benefit from ISAC mobile networks – and especially dedicated railway deployments – can be used to detect individuals, animals, or obstacles on train tracks and relay this information to train guidance systems to take counter-measures.

ISAC for industry and logistics

Modern industries and logistic operations often use automated guided vehicles (AGVs) and autonomous mobile robots (AMRs). At the same time, more and more industries rely on private wireless networks deployed to cater to the special needs of industrial and logistical operations. ISAC can be deployed using these networks to assist in AGV/AMR navigation but also for collision avoidance between AGVs, AMRs, and humans.

Immersive sensing

A more futuristic use case is immersive sensing. Here, ISAC helps build a digital representation of the physical world, a so-called digital twin. For mobile networks, a digital twin represents the physical world and can be used to simulate and predict network performance. A digital twin is obviously not limited to mobile networks, but can represent any physical world that should be digitally emulated. An example would be an extended reality game where the physical environment is incorporated by its digital twin into the game.

ISAC for mobile network performance

ISAC can also be used to improve the performance of the mobile network itself. For example, it is well known that millimeter wave (mmWave) communication performs best when it has Line of Sight (LoS) between transmitter and receiver. ISAC can be used to proactively detect a potential LoS interruption – for example due to an approaching vehicle – and enable the network to proactively switch the beam used to communicate with the receiver or hand it over to another transmission point.

Different levels of integration

The word “integrated” in integrated sensing and communication can mean different levels of integration, see Figure 1.

Figure 1: Different integration levels in an ISAC system

At the lowest integration levels, sites are shared between communication and sensing, but neither spectrum nor hardware deployed at the sites are shared. The next deeper integration levels share spectrum between communication and sensing, followed by sharing of hardware deployed at the site.

We believe sharing of sites, spectrum, and hardware are the most important integration steps since they allow leveraging available communication infrastructure for sensing.

Even deeper integration uses the same waveform for sensing and communication and reuses communication signals for sensing. We believe 6G will be based on orthogonal frequency division multiplexing (OFDM), therefore, sensing in 6G should be based on OFDM as well. Waveforms other than OFDM that would show superior performance and can be transmitted and received with hardware deployed for communication, could be considered.

The deepest level of integration involves reusing signals transmitted for communication. While this provides benefits of reduced overhead and energy consumption, it will only work in situations when an already transmitted communication signal has the correct properties for sensing (transmitted into the correct direction with sufficient bandwidth, etc.). We therefore believe this deepest level of integration is an optimization left for later stages of ISAC.

Above, we discussed the integration of ISAC from a deployment and radio perspective. To make ISAC truly useful it must be integrated into the network platform and provide an exposure interface to external applications and services. The exposure API would advertise the sensing capabilities of the network, receive sensing requests, and provide sensing results. This enables app developers to innovate new services that rely on sensing.

Sensing signal and processing

We propose to base ISAC on the principles of a pulse-Doppler radar due to its easy integration into the OFDM resource grid for communication, see Figure 2. Matched filtering of each received pulse is performed to obtain a correlation peak, which indicates the range to the target. Range resolution and accuracy improve with bandwidth.

Figure 2: A pulse-Doppler-based radar system can easily be integrated into an OFDM communication system.

Due to target movements, the propagation distance transmitter-target-receiver changes, that is, the phase of received pulses varies according to the distance change and thus target velocity. By comparing the phase between consecutively received sensing pulses, the target velocity can be determined. The closer two pulses occur in time, the higher the velocity that can be unambiguously estimated. The velocity resolution improves with the total duration of the pulse train.

The processing steps of matched filtering and phase comparison to determine range and velocity are typically done jointly using a 2D Fast Fourier Transform (FFT) which transforms the received pulse train into the delay (range) – Doppler (velocity) plane.

The direction of the target relative to the receiver position is determined by multi-antenna processing such as receiver beam forming and angle-of-arrival estimation. Mobile communication systems typically have very powerful antenna systems making them ideal for ISAC systems. The target's location is determined by combining its direction and range.

Spectrum for ISAC

Figure 3 shows all frequency bands we expect will become available in the 2030 timeframe and 6G will be able to reuse all existing 5G frequencies..

Figure 3: Multiple spectrum layers

Frequency range FR1 (0.41 GHz to 7.125 GHz) provides wide area coverage and midband time division duplex (TDD) spectrum channel bandwidths up to 100 MHz are available. Midband TDD is typically deployed with advanced antenna systems (AAS) with more than a hundred antenna elements. Large bandwidths and antenna arrays provide good range and angular resolution, respectively, important properties for ISAC applications. Channel bandwidths in lowband frequency division duplex (FDD) bands are often limited to 20 MHz and less advanced antenna systems are used, making these bands less attractive for ISAC applications requiring good resolution.

The mmWave spectrum (also called frequency range FR2, 24.25 GHz to 71.0 GHz) provides bandwidths of several 100 MHz. Its small wavelength allows for the use of electrical large antennas with several hundred antenna elements. These two factors make the mmWave spectrum attractive for sensing applications requiring superior resolution. The mmWave deployments provide less coverage and are typically deployed in hotspots or indoors. Therefore, ISAC applications using these frequencies can only be offered over a limited coverage area. High resolution together with local coverage make the mmWave spectrum an interesting band for industrial ISAC applications.

The centimeter wave (cmWave) spectrum from 7-15 GHz is an important potential spectrum addition for 6G. It provides wide bandwidths and, when deployed with large antenna arrays, it enables reuse of the midband deployment grid; in essence, mmWave-like bandwidths combined with midband coverage. It supports use cases requiring superior resolution over wide coverage areas.

We do not expect that sub-THz spectrum (100 GHz and above) will be introduced early on in 6G and therefore it is not further considered here. For more information on sensing in this frequency range, please refer to our earlier blog post.

Sensing topologies

The traditional radar topology is monostatic where the same node is used as transmitter and receiver, see Figure 4a. Depending on the distance (delay) to the target, the sensing signal transmission might still be ongoing when the earliest echo arrives. In an ISAC system, we typically want to detect close targets. With a distance of 50 m, for example, to the target, the reflected signal arrives 333 ns later. This can be compared with the OFDM symbol duration of 33 µs used in midband TDD deployments. The ongoing transmission creates strong self-interference during the reception of the weak target reflection. To cope with strong self-interference, the receiver must be capable of full-duplex operation, requiring complex radio and antenna solutions. Alternatively, a short sensing pulse combined with rapid transmit-receive switching ensures sensing signal transmission concludes before the target echo arrives. We believe both full-duplex and fast transmit-receive switches are more suitable for FR2 than FR1, making monostatic ISAC more attractive for FR2 than FR1.

Figure 4b shows a bistatic sensing system that avoids self-interference by using different nodes as transmitters and receivers. Accurate target location and velocity estimation require tight time- and frequency synchronization between transmitter and receiver, typically much tighter than what is needed for communication. To avoid much tighter per-node synchronization than what is needed for communication, we propose to use over-the-air synchronization: The receiver listens to a signal received over a reference path (preferably an LoS or stable and well-characterized non-LoS path) and obtains synchronization from this signal. A similar method is radio-interface based synchronization, which is already used today to synchronize base stations over the air.

Multistatic sensing is an extension to bistatic sensing which combines multiple bistatic sensing links, see Figure 4c.

Figure 4: a) Monostatic sensing topology, b) bistatic sensing topology, and c) multistatic sensing topology

Challenges

Classical sensing requires LoS to the target. The likelihood of LoS depends on the environment and target position. While it is usually abundantly available for aerial targets such as UAVs, it can pose a challenge for ground targets: Deployments of mobile networks minimize LoS to multiple transmission and reception points but even the likelihood of a single LoS link can be low due to environmental obstacles. Integration of end-user devices (early on, limited to operator-deployed devices), deployment of dedicated sensing receivers, densification of the network, and potentially advanced non-LoS sensing algorithms can mitigate this challenge.

Monostatic and bistatic sensing with base stations should use downlink slots in TDD to minimize interference to communication: From a device perspective, an interfering sensing transmission is like any other interfering downlink communication signal. However, the receiving base station needs to receive the weak target reflection when the network is in a downlink phase, that is, cross-link interference from adjacent carriers/sectors/sites/operator occurs. This interference can be very severe and requires mitigation schemes such as coordination between transmitters.

Privacy considerations

The current 5G user privacy solutions are generally designed around a known UE identifier whereas in sensing the target is an ordinary object located in a physical space. Such an object may be a data subject for which data collection and processing regulations may apply but does not have the means to express their preferences in contrast with the traditional communication subject. And together with challenges brought by the stringent rules and requirements for obtaining and maintaining consent for large-scale indirectly obtained data, we find consent as a legal basis for sensing infeasible at this point. Rather, we recommend focusing on use cases leveraging other legal basis, such as public safety, where personal data are being handled.

Use case example

In the following, we provide an example of how a multistatic ISAC system can be used to detect UAVs. The inter-site distance is 400 m and the network operates at 3.5 GHz. The UAVs are assumed to have an altitude of 300 m and a radar cross-section of -17 dBsqm. We assume cross-link interference is mitigated via coordination. Figure 5 shows a median location accuracy of approximately 2 m can be achieved. Per bistatic link, Cramer Rao bounds are used to determine SNR-dependent range and angle standard deviations. These estimates are then fused to obtain location values. To compensate for the “ideality” of the Cramer Rao bound, an implementation margin is subtracted from the bistatic SNR values.

Figure 5: CDF of achievable location accuracy for UAVs.

Summary and outlook

ISAC is an exciting new technology that has received considerable attention in academia. Initial standardization of ISAC is now also underway, with the completed use cases and requirements study by the 3GPP technical specification group Service and System Aspects, and the ongoing ISAC channel modeling study by 3GPP technical specification group Radio Access Network (RAN).

We see technology potential in ISAC, but it is important to recognize that the ISAC journey is still in its early stages and several challenges need to be addressed before its full realization. We expect the first standardized version of ISAC to be part of the initial 6G release in Rel-21.

Read more

What is 6G?

White paper: 6G – Connecting a cyber-physical world

White paper: 6G spectrum - enabling the future mobile life beyond 2030