Angle Estimation for Initial Access in True-Time-Delay Systems: Communication and Sensing Tradeoff
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Electrical Engineering. The day in electrical engineering: signals, systems and the hardware that carries them.
Rosa: We're looking at the paper: "Angle Estimation for Initial Access in True-Time-Delay Systems".
Dev: Here's the gist of the paper. In joint communication & sensing (JCNS) the data signal is used to obtain parameter estimates, such as the angle of arrival (AoA), which can be used to solve initial access problems in beam selection The work investigates the tradeoff between angle estimation accuracy and achievable data rate for uplink signals in a true-time-delay (TTD) array System Model Description The considered system configuration illustrates a transmitter radiating a signal and a ULA receiving it at multiple elements The received signal at the m-th antenna element experiences a channel gain based on the direction vector The baseband signal delayed by true-time delay is expressed in terms of frequency domain components The system function H(f, ϕ) can be expressed as a summation over the array elements Sensing Limit Derivation The Cramer-Rao bound provides a lower bound on the variance of any unbiased estimator and serves as a benchmark for evaluating angle estimation methods in TTD systems The Fisher information (FI) is defined as the negative second derivative of the log-likelihood function with respect to the angle Under assumptions of a flat spectrum and narrowband conditions, an approximation for the CRLB is given by an inequality involving signal energy and angular terms Waveform Design Optimization The design problem is formulated as maximizing the Fisher information under a data rate constraint This optimization is achieved by introducing a cost function Jn(γn) that balances sensing information, communication rate, and power consumption The resulting optimal power allocation is found via waterfilling where the water level depends on the sensing weight of each frequency bin Performance Evaluation Numerical simulations show that the derived BCRLB can be attained by a maximum likelihood (ML) estimator above a certain signal-to-noise ratio (SNR) threshold The analysis reveals that smaller prior widths result in higher accuracy in angle estimates, and the modified waterfilling approach yields power allocations optimal for maximizing sensing accuracy for a given data rate The ML estimator is shown to attain the bound up to specific SNR thresholds, validating the waveform design The Peak estimator is limited in angular resolution and does not attain the BCRLB, but its MSE is sufficient for beam selection in initial access The main motivation of demonstrating that the FI underlying the waveform design and consecutive sensing/communication tradeoff (27) is attainable, is verified by a ML estimator The work concludes that maximizing the Fisher information under a rate constraint leads to a convex problem solved by waterfilling across frequency bins The FI gain from waterfilling is inversely proportional to the prior angular width, with an example of 7.
Title and authors: Rosa: We’re looking at the paper titled "Angle Estimation for Initial Access in True-Time-Delay Systems: Communication and Sensing Tradeoff" by Carl Collmann, Ahmad Nimr, and Gerhard Fettweis. Rosa, this setup is about using the data signal to get parameter estimates like the angle of arrival.
Dev: That's right. It tackles a problem in joint communication and sensing where you use the data signal to obtain those parameter estimates, such as the angle of arrival, which is key for solving initial access problems when selecting a beam.
Rosa: The paper points out that true-time delay arrays are interesting here because their frequency dependent response encodes the angle information directly in the received signal's power spectrum. It’s like the system is inherently designed to hold that spatial data.
Taro: So, if we think about this outside of a lab setting, it suggests we can leverage existing communication infrastructure for sensing purposes without needing completely separate sensing equipment.
Dev: Exactly. The authors are investigating the trade-off between how accurate that angle estimation is versus the actual data rate you can achieve for uplink signals in these true-time delay arrays. It’s a fundamental tension they're exploring.
Rosa: They use the Fisher information to find the Bayesian Cramer-Rao lower bound, which is this theoretical minimum variance you can expect from any unbiased estimator when trying to estimate the angle.
Taro: And they then build up on that by showing how maximizing this Fisher information while keeping a data rate constraint leads them to an optimized power allocation strategy across different frequency bins.
Dev: That optimization is achieved by using waterfilling power per frequency bin dependent on sensing weights, which links the sensing quality directly into the power distribution decision.
Rosa: So, it’s not just about sending more bits; it's about designing the signal transmission itself to get better angle estimates for a given data rate, or at least to make them as good as possible under fixed power.
Taro: I wonder if this works well when the environment gets messy, like when you’re not in a controlled lab setting, and how resilient that design is.
Dev: That’s where we need to look at the performance evaluation, because simulations show that this derived Bayesian Cramer-Rao lower bound can be attained by a maximum likelihood estimator above a certain signal-to-noise ratio threshold.
Rosa: That SNR threshold is important because it sets the realistic conditions for when we can expect this high accuracy in angle estimation to actually happen.
Taro: So, if we're deploying this, we need to know what kind of noise level we're dealing with before we can trust the accuracy of the AoA estimate.
The paper's summary: Rosa: Moving into the actual summary of "Angle Estimation for Initial Access in True-Time-Delay Systems: Communication and Sensing Tradeoff," they lay out that this work investigates how to design a signal transmission that balances two conflicting goals.
Dev: The two conflicting goals are achieving a high angle estimation accuracy and maintaining a sufficient achievable data rate for the uplink signals. It’s about finding that sweet spot where neither one completely sacrifices the other.
Rosa: They set up the system model by describing how a transmitter radiates a signal and how multiple elements in an array, say M elements spaced by distance d, receive it, including common channel delay and propagation delays related to each element.
Taro: So they’re mapping out the physics of the ULA receiver setup first before diving into the information theory part? That sounds like a solid way to build up the technical foundation.
Dev: They do. They express the received signal in terms of frequency domain components, showing how it depends on both the frequency and that angle information you're trying to estimate. They define that system function H(f, phi) as a summation over all array elements.
Rosa: Then they apply some calculus to find the Fisher information for the angle, which gives them a lower bound on the variance of any unbiased estimator, which is this Bayesian Cramer-Rao lower bound.
Dev: That approximation involves an integral over some interval related to the frequency bins and terms involving signal energy and angular factors. They show how this Fisher information scales with things like signal energy Es and constants related to the angle.
Taro: So they're using those scaling laws to argue that as we change certain parameters, we can predict how much better our angle estimation will get without needing to run massive simulations every time.
Rosa: They then formulate the design problem as maximizing that Fisher information under a data rate constraint, which leads directly into the waterfilling power allocation per frequency bin dependent on sensing weights.
Dev: The core contribution is showing that this maximization results in a convex problem solved by waterfilling across the frequency bins, and they show how much better it is compared to using a flat spectrum signal.
Taro: So they’re suggesting that for uplink signals in these TTD arrays, we should move away from simple frequency flat signals towards this sensing-aware power allocation strategy.
The paper's improvements: Rosa: Now let's talk about the specific improvements the authors suggest to their initial problem, which is how they get that Fisher information and solve the trade-off.
Dev: They propose introducing a cost function, let's call it J n of gamma n, which is this function that balances sensing information gain, communication rate, and power consumption. This function guides the optimization process.
Rosa: That cost function is what allows them to solve the problem by waterfilling power per frequency bin dependent on those sensing weights. It’s a clever way to link the sensing objective with the communication objective.
Taro: So this means we're not just optimizing for one thing in isolation; we're making sure that every power decision has a dual purpose—doing both sensing and communicating effectively at once.
Dev: That’s right. The resulting optimal power allocation is found through waterfilling, and the water level in the waterfilling depends on the sensing weight of each frequency bin, which is crucial here.
Rosa: And they demonstrate that this modified waterfilling approach yields power allocations that are optimal for maximizing sensing accuracy for a given data rate. It’s tailored specifically to maximize sensing performance under those constraints.
Taro: So if we're deploying this, it means we can dynamically adjust the transmit power per frequency bin based on those sensing weights derived from the waterfilling solution, ensuring that bins carrying high angle information get more power than in classic waterfilling scenarios.
Dev: That’s a key distinction; it ensures that bins with high angle sensitivity are prioritized in terms of power distribution. It’s not just generic communication optimization anymore; it's sensing-informed communication optimization.
Rosa: And they confirm that this approach is attainable by a maximum likelihood estimator above specific SNR thresholds, which validates the whole waveform design when you look at practical implementation.
Conclusion: Dev: So we’re wrapping up the discussion on "Angle Estimation for Initial Access in True-Time-Delay Systems: Communication and Sensing Tradeoff." To summarize, the main finding is that maximizing Fisher information under a data rate constraint leads to a convex problem solved by waterfilling across frequency bins.
Rosa: And they highlight that the gain in Fisher information you get from using this waterfilling approach is inversely proportional to your prior angular width, which they give an example of being about seven times more information.
Taro: That inverse relationship with the prior width is a really important detail for us when thinking about how we set our initial guess for the angle.
Dev: And they confirm that this method gives us a concrete way to design uplink signals that can simultaneously communicate and sense, which pushes what we can do with true-time delay arrays.
Rosa: It shows that the underlying Fisher information is attainable by a maximum likelihood estimator above certain signal-to-noise ratio thresholds, which confirms the feasibility of the whole thing in practice.
Dev: So this paper gives us a solid framework for designing next generation systems where we can optimize power allocation based on sensing requirements, and we can move on to looking at how these things fail when they don't behave as expected.
Taro: I just keep thinking about what happens when the world misbehaves, like if the spectrum isn't flat, because that’s the next big hurdle for real deployment.
Rosa: Right. That’s where we’ll take it next time.
Technische Universitat Dresden
eess.SP
Submitted: 2026-11-05
Updated: 2026-11-05
Comments: 6 pages, 5 figures, 1 table
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 74/100
The gist: In joint communication & sensing (JCNS) the data signal is used to obtain parameter estimates, such as the angle of arrival (AoA)…
Key concepts
- Joint Communication & Sensing (JCNS)
- This involves using a single data signal simultaneously for two tasks: communicating information and sensing the environment. In this context, the data itself is used to estimate parameters like the angle of arrival (AoA) of a signal, which is crucial for selecting the best communication beam.
- Cramer-Rao Bound (CRLB)
- The CRLB sets a theoretical minimum variance for any unbiased estimator used to estimate an unknown parameter, such as the AoA. It acts as a benchmark to measure how accurate an angle estimation method is. The paper uses this bound to evaluate the performance of their proposed waveform designs.
- Fisher Information (FI)
- Fisher Information quantifies how much information a signal carries about an unknown parameter, like the angle. Maximizing this information under a data rate constraint guides the design of optimal waveforms that provide the best possible trade-off between sensing quality and communication speed.
Terminology
Summary
In joint communication & sensing (JCNS) the data signal is used to obtain parameter estimates, such as the angle of arrival (AoA), which can be used to solve initial access problems in beam selection The work investigates the tradeoff between angle estimation accuracy and achievable data rate for uplink signals in a true-time-delay (TTD) array
System Model Description
The considered system configuration illustrates a transmitter radiating a signal and a ULA receiving it at multiple elements The received signal at the m-th antenna element experiences a channel gain based on the direction vector The baseband signal delayed by true-time delay is expressed in terms of frequency domain components The system function H(f, ϕ) can be expressed as a summation over the array elements
Sensing Limit Derivation
The Cramer-Rao bound provides a lower bound on the variance of any unbiased estimator and serves as a benchmark for evaluating angle estimation methods in TTD systems The Fisher information (FI) is defined as the negative second derivative of the log-likelihood function with respect to the angle Under assumptions of a flat spectrum and narrowband conditions, an approximation for the CRLB is given by an inequality involving signal energy and angular terms
Waveform Design Optimization
The design problem is formulated as maximizing the Fisher information under a data rate constraint This optimization is achieved by introducing a cost function Jn(γn) that balances sensing information, communication rate, and power consumption The resulting optimal power allocation is found via waterfilling where the water level depends on the sensing weight of each frequency bin
Performance Evaluation
Numerical simulations show that the derived BCRLB can be attained by a maximum likelihood (ML) estimator above a certain signal-to-noise ratio (SNR) threshold The analysis reveals that smaller prior widths result in higher accuracy in angle estimates, and the modified waterfilling approach yields power allocations optimal for maximizing sensing accuracy for a given data rate The ML estimator is shown to attain the bound up to specific SNR thresholds, validating the waveform design The Peak estimator is limited in angular resolution and does not attain the BCRLB, but its MSE is sufficient for beam selection in initial access The main motivation of demonstrating that the FI underlying the waveform design and consecutive sensing/communication tradeoff (27) is attainable, is verified by a ML estimator
Conclusion
The work concludes that maximizing the Fisher information under a rate constraint leads to a convex problem solved by waterfilling across frequency bins The FI gain from waterfilling is inversely proportional to the prior angular width, with an example of 7.
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Waterfilling based waveform design for uplink signals allows base stations to optimize power allocation across frequency bins by maximizing Fisher information while maintaining a minimum data rate, leading to improved angle estimation accuracy under a fixed power constraint.
-
The improved AI system can perform joint communication and sensing (JCNS) tasks where the data signal is used to obtain parameter estimates like the angle of arrival (AoA), enabling smarter context-aware mobile communication systems and accurate beam selection.
-
The system can achieve a Bayesian Cramer-Rao lower bound that is attainable by a maximum likelihood estimator above a certain signal-to-noise ratio threshold, confirming that the Fisher information underlying the waveform design is achievable in practice.
-
The AI system can utilize a two-stage sensing scheme: first using a frequency flat waveform for coarse estimation to create a narrow prior, and then employing waterfilling optimized waveforms for the second stage to maximize angle estimation accuracy.
-
The system can dynamically adjust its transmit power per frequency bin based on sensing weights derived from the waterfilling solution, ensuring that bins carrying high angle information receive more power than in classic waterfilling scenarios.
Abstract
In joint communication & sensing (JCNS) the data signal is used to obtain parameter estimates. An example is the angle of arrival (AoA) which can be used to solve the initial access problem of beam selection. true-time delay (TTD) arrays are attractive for this purpose, as their frequency-dependent response encodes the angle information in the receive signal power spectrum. This work investigates the tradeoff between angle estimation accuracy and achievable data rate for uplink signals in a TTD array. The Fisher information (FI) for the angle is used to find the bayesian Cramer-Rao lower bound (BCRLB) under the assumption of a raised cosine distributed angle prior. Maximization of the FI under a data rate constraint is solved by waterfilling power per frequency bin dependent on sensing weights. Compared to a frequency flat signal, the optimized power allocation increases the FI about 4.9 dB at 5% rate loss for prior width of plus or minus 10°. Numerical simulations show that the derived BCRLB can be attained by a maximum likelihood (ML) estimator above an signal-to-noise ratio (SNR) threshold, confirming that the FI underlying the waveform design is attainable.
Sources
Related papers
- Open-H-RF: A Large-Scale Ultrasound Raw Channel Dataset Initiative
- Scalable Attention for 5G NR Channel Estimation
- Low-Altitude Wireless Networks: A Comprehensive Survey
- Sensing-Assisted OFDM with Index Modulation: A Waveform for Optical Wireless Integrated Communication and Localisation
- Iteration-Free Candidate Pruning and an Index-Domain VLSI Architecture for Sort-Free MIMO Box Decoding up to 1024-QAM
- A Foundation Model for Large-Scale Wireless Network Planning, Operation and Optimization