Zhuoran Wang

Acquiring and processing full‐motion details in machine vision typically consumes a substantial amount of energy. In contrast, a hierarchical processing architecture, combining a low‐power standby front end with an on‐demand activated back end, provides an optimized energy‐performance tradeoff. To achieve this, the complete acquisition and decoupling of static (brightness) and dynamic (amplitude and polarity) output at the sensory level are essential for activating on‐demand vision function. Here, we report a differential image sensor (DIS) that leverages differential photodiodes with decoupled differential and tunneling modes. These modes can be read out via conventional ROICs, paving the way for the up‐scaled integration (e.g., 640 × 512). With on‐demand activated dynamic and static modes, the DIS implements a hierarchical motion‐processing pipeline—from sparse motion detection to optical flow and depth analysis. This work provides a power‐efficient and scalable strategy for advancing vision‐based AIoT applications.

The rise of the Artificial Intelligence of Things (AIoT) demands sensory systems with reduced size, weight, and power (SWaP). The processing‐in‐sensor (PIS) paradigm offers a solution, providing superior compactness and power‐efficiency, critical for edge vision applications. Among emerging optoelectronic neuromorphic devices, the direct photocurrent computing (DPC) route is uniquely attractive, using photoresponsivity to encode weights for in‐sensor multiply–accumulate (MAC) operations. However, current DPC devices rely on electrical signals for weight programming, which complicates circuitry and limits bandwidth compared to all‐optical approaches. To address this, we present an optically programmable DPC device based on a vacancy‐modulated bismuth oxyselenide (BOS) material platform. Critically, the reversible surface hydroxyl dissociation is found to reconfigure oxygen vacancy dynamics upon ultraviolet light, enabling the spectrally decoupled weight programming and photocurrent computing. Based on this, we demonstrate a BOS array implemented PIS hardware for low‐power, coarse classification and as a pre‐processing unit for more complex vision tasks in a processing‐near‐sensor (PNS) paradigm. Finally, a hybrid architecture is proposed to intelligently allocate computational resources between PIS and PNS, promising for an optimal balance of power and performance for next‐generation edge AIoT applications.

With the rapid development of micro/nanofabrication technologies, the concept of transformable kirigami has been applied for device fabrication in the microscopic world. However, most nano-kirigami structures and devices were typically fabricated or transformed at fixed positions and restricted to limited mechanical motion along a single axis due to their small sizes, which significantly limits their functionalities and applications. Here, we demonstrate the precise shaping and position control of nano-kirigami microrotors. Metallic microrotors with size of ~10 micrometers were deliberately released from the substrates and readily manipulated through the multimode actuation with controllable speed and direction using an advanced optoelectronic tweezers technique. The underlying mechanisms of versatile interactions between the microrotors and electric field are uncovered by theoretical modeling and systematic analysis. This work reports a novel methodology to fabricate and manipulate micro/nanorotors with well-designed and sophisticated kirigami morphologies, providing new solutions for future advanced optoelectronic micro/nanomachinery.

An innovative, low-temperature, post-liftoff-free, seed-confined fabrication strategy is developed to achieve designated patterns of flower-shaped AgBiS ₂ nanostructures on rigid or flexible substrates for broadband and eco-friendly photodetectors.

The Conflict-Based Search (CBS) algorithm is one of the Multi-Agent Path Finding (MAPF) algorithms that has received much attention in recent years. It is a two-level algorithm. The high-level detects and resolves conflicts on a Constraint Tree (CT). Its low-level algorithm is to plan the optimal path for a given agent that satisfies all the constraints of the agent. In this paper, the Space-Time A* (STA*) algorithm is used as the low-level algorithm, which adds the time dimension to the A* algorithm. It means that the planning path for a single agent tends to generate and extend more and more sub-nodes. As the complexity of the problem increases, the efficiency of the low-level solution becomes an increasingly large part of the overall solution efficiency. This leads to an inefficient overall solution. To address this problem, an improved CBS algorithm is proposed. This algorithm is used when the current node in CT has conflicts. It uses current node information and Multi-value Decision Diagram (MDD) technology to pre-plan before formal re-planning. The pre-planning phase entails obtaining information pertaining to the path that satisfies the new constraint, which is referred to as pre-planning information. During the formal re-planning, the pre-planning information is used to guide the low-level algorithm to select the generated sub-nodes. The simulation experiments show that CBS with pre-planning improves the solution time by 24.4% and reduces the search space by 38.7%.