Deep Learning for ISAC-Enabled End-to-End Predictive Beamforming in Vehicular Networks
Bibliographic record
Abstract
Integrated sensing and communications (ISAC) has emerged as a promising technology for predictive beamforming design in vehicle-to-infrastructure (V2I) networks. Most of the existing works use a two-step approach for predictive beamforming design. The first step is to estimate the state parameters of a vehicle (e.g., angle of arrival (AoA), channel state information (CSI)) from the received sensing signal samples at the road side unit (RSU). The second step is to determine the beamforming vector based on the estimated parameters. However, estimation errors may be introduced in the first step which impacts the subsequent beamforming design and leads to degradation in the achievable rate. In this work, by using deep learning, we propose an ISAC-enabled end-to-end predictive beamforming (E2E-PB) approach to obtain the beamforming vector directly from the reflected signal samples. The proposed approach does not require an intermediate state parameters estimation step. We develop an attention-based long short-term memory (LSTM) network to capture the temporal correlation in the reflected signal samples and determine the beamformer. The network is trained in an unsupervised manner to maximize the achievable rate. We compare our proposed E2E-PB approach with two state-of-the-art schemes, namely, the extended Kalman filtering framework and a deep learning based two-step approach. The results show that our proposed E2E-PB approach obtains a higher achievable rate than the other two baseline schemes, and has close performance when compared with the optimal beamforming design with perfect CSI.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.002 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".