Lifeng Zhou

Hyperspectral imaging (HSI) has emerged as a promising technique for microplastic detection through analysis of reflectance variations across multiple wavelengths. Traditional approaches have focused primarily on isolated microplastic particles, requiring labor‐intensive separation procedures impractical for routine monitoring. The challenge of detecting microplastics directly on food surfaces stems from spectral similarities between microplastics and food matrices, making differentiation difficult using conventional methods. Leveraging recent advances in machine learning, this study explores how artificial neural networks (ANN) and one‐dimensional convolutional neural networks (1D‐CNN) can identify subtle spectral differences to detect microplastic particles on seafood without isolation. We systematically evaluated model architectures, preprocessing techniques, and hyperparameter configurations to optimize detection performance using hyperspectral data from tilapia samples contaminated with polyethylene microspheres. Our findings demonstrate that 1D‐CNN models trained on hyperspectral data without dimensionality reduction significantly outperform other approaches, achieving object‐level detection F1 scores of 0.963 for 600‐μm particles and 0.950 for 300‐μm particles. This detection strategy represents a substantial improvement over traditional methods and highlights the potential of deep learning–based approaches for non‐destructive, efficient microplastic detection in food safety applications.

•Developed intelligent food portioning system for small-scale operations•π0 model achieved 100% success rate with only 30-50 training demonstrations•System adapts across diverse food types: shrimp, grapes, and garlic cloves•Vision-language-action models integrated with real-time weight monitoring•Efficient operation: 15.23 seconds to portion 30g of shrimp with high accuracy
The food industry faces significant barriers to adopting automation, as most of the food service, retail, and processing operations are small businesses that lack the financial capacity to invest in conventional industrial automation systems. Food portioning is a fundamental operation across food industry sectors, yet it remains highly labor-intensive for small businesses. Existing automated portioning systems are generally designed for single-product, large-scale processing, rendering them financially prohibitive and operationally inflexible for small-scale operations with diverse product requirements. Advancements in artificial intelligence (AI) provide promising avenues for the development of cost-effective automation systems for small food businesses, offering adaptable solutions capable of handling multiple food types with flexibility and precision. This study proposes an AI-driven, low-cost food portioning framework as a proof-of-concept solution that integrates weight sensing with vision-language-action (VLA) control to enable adaptable handling of diverse food products. The system employs You-Only-Look-Once (YOLO)-based vision models to interpret digital scale readings while coordinating robotic picking mechanisms that transfer food items until the target weight is reached. Three vision-language models, namely Action Chunking with Transformers (ACT), OpenVLA with Optimized Fine-Tuning (OpenVLA-OFT), and π0, were evaluated on shrimp (30g), grapes (50g), and garlic (20g), demonstrating adaptability across diverse food types. The π0 model achieved a 100% success rate using only 30–50 demonstrations per food type and demonstrated efficient operational performance (e.g., 15.23 seconds to portion 30 g of shrimp). This framework demonstrates the potential for adaptive automation in small-scale food businesses, providing a preliminary foundation that addresses single-product automation limitations in food packaging, distribution and service operations.
[Display omitted]

In this article, we study the problem of game-theoretic robot allocation where two players strategically allocate robots to compete for multiple sites of interest. Robots possess offensive or defensive capabilities to interfere and weaken their opponents to take over a competing site. This problem belongs to the conventional an acronym colonel blotto game (CBG). Considering the robots' heterogeneous capabilities and environmental factors, we generalize the conventional Blotto game by incorporating heterogeneous robot types and graph constraints that capture the robot transitions between sites. Then, we employ the double oracle algorithm (DOA) to solve for the Nash equilibrium of the generalized Blotto game. Particularly, for cyclic-dominance-heterogeneous (CDH) robots that inhibit each other, we define a new transformation rule between any two robot types. Building on the transformation, we design a novel utility function to measure the game's outcome quantitatively. Moreover, we rigorously prove the correctness of the designed utility function. Finally, we conduct extensive simulations to demonstrate the effectiveness of DOA on computing Nash equilibrium for homogeneous, linear heterogeneous, and CDH robot allocation on graphs.

In this work, we study the problem of decentralized multi-agent perimeter defense that asks for computing actions for defenders with local perceptions and communications to maximize the capture of intruders. One major challenge for practical implementations is to make perimeter defense strategies scalable for large-scale problem instances. To this end, we leverage graph neural networks (GNNs) to develop an imitation learning framework that learns a mapping from defenders’ local perceptions and their communication graph to their actions. The proposed GNN-based learning network is trained by imitating a centralized expert algorithm such that the learned actions are close to that generated by the expert algorithm. We demonstrate that our proposed network performs closer to the expert algorithm and is superior to other baseline algorithms by capturing more intruders. Our GNN-based network is trained at a small scale and can be generalized to large-scale cases. We run perimeter defense games in scenarios with different team sizes and configurations to demonstrate the performance of the learned network.

We study the problem of incorporating risk while making combinatorial decisions under uncertainty. We formulate a discrete submodular maximization problem for selecting a set using conditional value at risk (CVaR), a risk metric commonly used in financial analysis. While the CVaR has recently been used in the optimization of linear cost functions in robotics, we take the first step toward extending this to discrete submodular optimization and provide several positive results. Specifically, we propose the sequential greedy algorithm that provides an approximation guarantee on finding the maxima of the CVaR cost function under a matroid constraint. The approximation guarantee shows that the solution produced by our algorithm is within a constant factor of the optimal and an additive term that depends on the optimal. Our analysis uses the curvature of the submodular set function and proves that the algorithm runs in polynomial time. This formulates a number of combinatorial optimization problems that appear in robotics. We use two such problems, i.e., vehicle assignment under uncertainty for mobility on demand and sensor selection with failures for environmental monitoring, as case studies to demonstrate the efficacy of our formulation. We also study the problem of adaptive risk-aware submodular maximization. We design a heuristic solution that triggers the replanning only when certain conditions are satisfied, to eliminate unnecessary planning. In particular, for the online mobility-on-demand study, we propose an adaptive triggering assignment algorithm that triggers a new assignment only when it can potentially reduce the waiting time at demand locations. We verify the performance of the proposed algorithms through simulations.

In this article, we design algorithms to protect swarm-robotics applications against sensor denial-of-service attacks on robots. We focus on applications requiring the robots to jointly select actions, e.g., which trajectory to follow, among a set of available actions. Such applications are central in large-scale robotic applications, such as multirobot motion planning for target tracking. But the current attack-robust algorithms are centralized. In this article, we propose a general-purpose distributed algorithm toward robust optimization at scale, with local communications only. We name it distributed robust maximization ( DRM ). DRM proposes a divide-and-conquer approach that distributively partitions the problem among cliques of robots. Then, the cliques optimize in parallel, independently of each other. We prove DRM achieves a close-to-optimal performance. We demonstrate DRM 's performance in Gazebo and MATLAB simulations, in scenarios of active target tracking with swarms of robots . In the simulations, DRM achieves computational speed-ups, being 1 to 2 orders faster than the centralized algorithms. Yet , it nearly matches the tracking performance of the centralized counterparts. Since, DRM overestimates the number of attacks in each clique, in this article, we also introduce an improved distributed robust maximization ( IDRM ) algorithm. IDRM infers the number of attacks in each clique less conservatively than DRM by leveraging three-hop neighboring communications. We verify IDRM improves DRM 's performance in simulations.

In this letter, we consider a distributed submodular maximization problem for multi-robot systems when attacked by adversaries. One of the major challenges for multi-robot systems is to increase resilience against failures or attacks. This is particularly important for distributed systems under attack as there is no central point of command that can detect, mitigate, and recover from attacks. Instead, a distributed multi-robot system must coordinate effectively to overcome adversarial attacks. In this work, our distributed submodular action selection problem models a broad set of scenarios where each robot in a multi-robot system has multiple action selections that may fulfill a global objective, such as exploration or target tracking. To increase resilience in this context, we propose a fully distributed algorithm to guide each robot's action selection when the system is attacked. The proposed algorithm guarantees performance in a worst-case scenario where up to a portion of the robots malfunction due to attacks. Importantly, the proposed algorithm is also consistent, as it is shown to converge to the same solution as a centralized method. Finally, a distributed resilient multi-robot exploration problem is presented to confirm the performance of the proposed algorithm.

Purpose of Review Deploying a team of robots that can carefully coordinate their actions can make the entire system robust to individual failures. In this report, we review recent algorithmic development in making multi-robot systems robust to environmental uncertainties, failures, and adversarial attacks. Recent Findings We find the following three trends in the recent research in the area of multi-robot coordination: (1) resilient coordination to either withstand failures and/or attack or recover from failures/attacks; (2) risk-aware coordination to manage the trade-off risk and reward, where the risk stems due to environmental uncertainty; (3) Graph neural networks based coordination to learn decentralized multi-robot coordination policies. These algorithms have been applied to tasks such as formation control, task assignment and scheduling, search and planning, and informative data collection. Summary In order for multi-robot systems to become practical, we need coordination algorithms that can scale to large teams of robots dealing with dynamically changing, failure-prone, contested, and uncertain environments. There has been significant recent research on multi-robot coordination that has contributed resilient and risk-aware algorithms to deal with these issues and reduce the gap between theory and practice. Learning-based approaches have been seen to be promising, especially since they can learn who, when, and how to communicate for effective coordination. However, these algorithms have also been shown to be vulnerable to adversarial attacks, and as such developing learning-based coordination strategies that are resilient to such attacks and robust to uncertainties is an important open area of research.