arXiv subjects

Xiaozhou Zhang

Publications and source records attributed to Xiaozhou Zhang.

5 recordsLinked to original sources

arXiv2026

CoLT-Drive: Counterfactual Long-Tail Benchmarking and Knowledge-Preserving Adaptation for Driving Affordance Prediction

Long-tail autonomous driving failures are often framed as rare-object recognition errors. We argue that this view is incomplete: the decision-critical question is not only whether a model recognizes an unusual object, but whether it infers how that object changes the ego vehicle's feasible high-level actions. We formalize this problem as decision-level driving affordance prediction, where a model maps a front-view image, ego-motion history, and navigation command to a structured longitudinal--lateral meta-action. To evaluate this capability, we introduce CoLT-Drive, a 3,536-sample counterfactual long-tail benchmark that inserts rare objects into otherwise fixed driving scenes and measures whether models predict acceptable action pairs. To improve deployable small VLMs, we propose KPA, a knowledge-preserving adaptation framework that combines structured perception-to-decision prompting, SLERP-based expert merging, and RegMoE, a regime-aware LoRA mixture-of-experts module. KPA preserves the pretrained model's open-world knowledge while allocating lightweight adaptation capacity to different driving decision regimes. Experiments on an in-domain driving split and CoLT-Drive show that KPA achieves 60.8\% pair accuracy on CoLT-Drive, outperforming the pretrained Qwen3-VL-2B baseline (50.3\%) and LoRA SFT (32.4\%) while maintaining competitive in-domain accuracy. Our benchmark and code are available at https://huggingface.co/datasets/tangzx2024/CoLT-Drive and https://github.com/tangzhengxu/CoLT-Drive.

cs.CV↗

arXiv2026

Beyond Textual Chain-of-Thought: A Survey on Action-Grounded Reasoning in Autonomous Driving

Chain-of-thought (CoT) reasoning powers generative models by eliciting intermediate steps before producing an answer. In autonomous driving, the answer is a continuous action. Thus its reasoning must share the same spatiotemporal structure as the physical world. This survey studies the resulting shift from textual CoT to action-grounded reasoning. Surveying 171 papers, including 130 method papers and 41 benchmarks, datasets, surveys, and analysis papers, we propose a representation-centered taxonomy that treats the form of the intermediate state as the organizing axis. We systematize the 130 methods into four categories: language-based, visual-spatial, latent-dynamic, and externalized reasoning, further divided into 13 subtypes tied to distinct regions of interests. Our synthesis shows that the open frontier of reasoning in driving agents lies in intermediate representations that can be grounded in the real world, coupled to real-time action, and verified under safety-critical systems. Project page: https://github.com/tangzhengxu/awesome-av-cot.

cs.CV↗

arXiv2024

Speeding Up Path Planning via Reinforcement Learning in MCTS for Automated Parking

In this paper, we address a method that integrates reinforcement learning into the Monte Carlo tree search to boost online path planning under fully observable environments for automated parking tasks. Sampling-based planning methods under high-dimensional space can be computationally expensive and time-consuming. State evaluation methods are useful by leveraging the prior knowledge into the search steps, making the process faster in a real-time system. Given the fact that automated parking tasks are often executed under complex environments, a solid but lightweight heuristic guidance is challenging to compose in a traditional analytical way. To overcome this limitation, we propose a reinforcement learning pipeline with a Monte Carlo tree search under the path planning framework. By iteratively learning the value of a state and the best action among samples from its previous cycle's outcomes, we are able to model a value estimator and a policy generator for given states. By doing that, we build up a balancing mechanism between exploration and exploitation, speeding up the path planning process while maintaining its quality without using human expert driver data.

cs.AI↗

arXiv2023

Multistatic Integrated Sensing and Communication System in Cellular Networks

A novel multistatic multiple-input multiple-output (MIMO) integrated sensing and communication (ISAC) system in cellular networks is proposed. It can make use of widespread base stations (BSs) to perform cooperative sensing in wide area. This system is important since the deployment of sensing function can be achieved based on the existing mobile communication networks at a low cost. In this system, orthogonal frequency division multiplexing (OFDM) signals transmitted from the central BS are received and processed by each of the neighboring BSs to estimate sensing object parameters. A joint data processing method is then introduced to derive the closed-form solution of objects position and velocity. Numerical simulation shows that the proposed multistatic system can improve the position and velocity estimation accuracy compared with monostatic and bistatic system, demonstrating the effectiveness and promise of implementing ISAC in the upcoming fifth generation advanced (5G-A) and sixth generation (6G) mobile networks.

eess.SP↗

arXiv2019

Max-Min Fairness User Scheduling and Power Allocation in Full-Duplex OFDMA Systems

In a full-duplex (FD) multi-user network, the system performance is not only limited by the self-interference but also by the co-channel interference due to the simultaneous uplink and downlink transmissions. Joint design of the uplink/downlink transmission direction of users and the power allocation is crucial for achieving high system performance in the FD multi-user network. In this paper, we investigate the joint uplink/downlink transmission direction assignment (TDA), user paring (UP) and power allocation problem for maximizing the system max-min fairness (MMF) rate in a FD multi-user orthogonal frequency division multiple access (OFDMA) system. The problem is formulated with a two-time-scale structure where the TDA and the UP variables are for optimizing a long-term MMF rate while the power allocation is for optimizing an instantaneous MMF rate during each channel coherence interval. We show that the studied joint MMF rate maximization problem is NP-hard in general. To obtain high-quality suboptimal solutions, we propose efficient methods based on simple relaxation and greedy rounding techniques. Simulation results are presented to show that the proposed algorithms are effective and achieve higher MMF rates than the existing heuristic methods.

cs.IT↗