Mixed traffic environments, comprising autonomous vehicles (AVs) and human-driven vehicles (HDVs), present substantial challenges for developing negotiation policies. These policies are essential for enabling AVs to make adaptive decisions and achieve harmonious interactions with HDVs in dynamic and complex scenarios. Despite their potential in autonomous driving, reinforcement learning-based decision-making schemes remain insufficient in addressing interaction-critical situations. To overcome this limitation, this study proposes a graph-enhanced negotiation-aware policy optimization (GNPO) framework, which embeds interaction awareness into all key components, including representation, understanding, and response. To support interaction representation, the spatio-temporal directed graph (STDG) captures the dynamic and asymmetric nature of interactions. It integrates multiple directed graph topologies, each encoding distinct social priors, to collectively depict realistic interactive behaviors. Additionally, a multimodal feature extractor fuses environmental perception, interaction cues, and task-related information, enabling comprehensive state understanding. Building on this representation, GNPO incorporates an interaction-critical reward function and an entropy-regularized adaptive policy optimization scheme, both designed to promote harmonious and context-aware behaviors in dense traffic scenarios. The proposed framework is validated in both interaction-intensive simulated environments and real-data-based digital twin scenarios, with its effectiveness quantitatively demonstrated by substantial improvements in success rate, safety, and efficiency. Qualitative evaluations further show that GNPO is capable of generating human-like cooperative and competitive behaviors.

Solving the sum rate maximization problem for interference reduction in multi-cell multi-user multiple-input multiple-output (MIMO) wireless communication systems has been investigated for a decade. Several machine learning-assisted methods have been proposed under conventional sum rate maximization frameworks, such as the Weighted Minimum Mean Square Error (WMMSE) framework. However, existing learning-assisted methods suffer from a deficiency in parallelization, and their performance is intrinsically bounded by WMMSE. In contrast, we propose a structural learning-only framework from the abstraction of WMMSE. Our proposed framework increases the solvability of the original MIMO sum rate maximization problem by dimension expansion via a unitary learnable parameter matrix to create an equivalent problem in a higher dimension. We then propose a structural solution updating method to solve the higher dimensional problem, utilizing neural networks to generate the learnable matrix-multiplication parameters. We show that the proposed structural learning framework achieves lower complexity than WMMSE thanks to its parallel implementation. Simulation results under practical communication network settings demonstrate that our proposed learning-only framework achieves up to 98% optimality over state-of-the-art algorithms while providing up to 47× acceleration in various scenarios.

Precoding design is essential for massive multi-user multiple-input multiple-output (MU-MIMO) systems, which aims at maximizing the weighted sum-rate (WSR). This problem is known to be NP-hard, and iterative algorithms are typically used to approximately solve it. The weighted minimum mean-squared error (WMMSE) algorithm is a popular solver for WSR maximization, which efficiently finds a local maxima of WSR. In this work, we introduce a novel extrapolation technique to further accelerate WMMSE. This technique is inspired by the momentum technique in convex optimization, and can be interpreted as an accelerated second-order method. The merits of the proposed extrapolation technique are (i) lightweight, as it almost does not increase the iteration complexity, (ii) generic, since it works in various settings such as the sum power constraint or per-antenna power constraint cases and coordinated multi-point joint transmission networks, and (iii) effective, that our simulation results show it significantly accelerates the convergence of WMMSE in the high channel correlation regime.

Accurate digital orthophoto map generation from high-resolution aerial images is important in various applications. Compared with the existing commercial software and the current state-of-the-art mosaicing systems, a novel fast georeferenced orthophoto mosaicing framework is proposed in this study. The framework can adapt to the challenging requirements of high-accuracy orthoimage generations with relatively fast speed, even if the overlap rate is low. We provide appearance and spatial correlation-constrained fast low-overlap neighbor candidate query and matching. On the basis of GPS information, we introduce an absolute position and rotation-averaging strategy for global pose initialization, which is essential for the high convergence and efficiency of nonconvex pose optimization of every image. We also propose a planar-restricted global pose graph optimization method. The optimization is extremely efficient and robust considering that point clouds are parameterized to planes. Finally, we apply a matching graph-based exposure compensation and region reduction algorithm for large-scale and high-resolution image fusion with high efficiency and novel precision. Experimental results demonstrate that our method can achieve the state-of-the-art performance while maintaining high precision and robustness.

Cellular network performance depends heavily on the configuration of its network parameters. Current practice of parameter configuration relies largely on expert experience, which is often suboptimal, time-consuming, and error-prone. Therefore, it is desirable to automate this process to improve the accuracy and efficiency via learning-based approaches. However, such approaches need to address several challenges in real operational networks: the lack of diverse historical data, a limited amount of experiment budget set by network operators, and highly complex and unknown network performance functions. To address those challenges, we propose a collaborative learning approach to leverage data from different cells to boost the learning efficiency and to improve network performance. Specifically, we formulate the problem as a transferable contextual bandit problem, and prove that by transfer learning, one could significantly reduce the regret bound. Based on the theoretical result, we further develop a practical algorithm that decomposes a cell's policy into a common homogeneous policy learned using all cells' data and a cell-specific policy that captures each individual cell's heterogeneous behavior. We evaluate our proposed algorithm via a simulator constructed using real network data and demonstrates faster convergence compared to baselines. More importantly, a live field test is also conducted on a real metropolitan cellular network consisting 1700+ cells to optimize five parameters for two weeks. Our proposed algorithm shows a significant performance improvement of 20%.