Publications
RoboLayout: Differentiable 3D Scene Generation for Embodied Agents
Recent advances in vision–language models (VLMs) have shown strong potential for spatial reasoning and 3D scene layout generation from open-ended language instructions. However, generating layouts that are not only semantically coherent but also feasible for interaction by embodied agents remains challenging, particularly in physically constrained indoor environments. In this paper, RoboLayout is introduced as an extension of LayoutVLM that augments the original framework with agent-aware reasoning and improved optimization stability. RoboLayout integrates explicit reachability constraints into a differentiable layout optimization process, enabling the generation of layouts that are navigable and actionable by embodied agents. Importantly, the agent abstraction is not limited to a specific robot platform and can represent diverse entities with distinct physical capabilities, such as service robots, warehouse robots, humans of different age groups, or animals, allowing environment design to be tailored to the intended agent. In addition, a local refinement stage is proposed that selectively re-optimizes problematic object placements while keeping the remainder of the scene fixed, improving convergence efficiency without increasing global optimization iterations. Overall, RoboLayout preserves the strong semantic alignment and physical plausibility of LayoutVLM while enhancing applicability to agent-centric indoor scene generation, as demonstrated by experimental results across diverse scene configurations.
SprintAgent: Agentic Construction of Execution-Ready Project Roadmaps
Autonomous agentic systems built with large language models (LLMs) introduce engineering trade-offs that are difficult to manage with conventional software infrastructure. This paper presents *SprintAgent*, an open-source framework designed as a production-ready scaffold for hierarchical agentic systems. The framework includes a planner–designer–critic orchestration loop, Hydra-based inversion-of-control configuration, SQLite-backed session management with turn trimming, checkpoint-based recovery, isolated parallel execution, and unified logging for observability. SprintAgent emphasizes modularity, allowing practitioners to extend or replace components with minimal friction. An agentic project-management system is implemented as a reference application to demonstrate the framework in practice.
Optimized Real-Time Soft Analyzer for Chemical Process Using Artificial Intelligence
This work explores AI-driven approaches for process monitoring and variable estimation in chemical industries, focusing on difficult-to-measure parameters. A multi-layer perceptron neural network is optimized and validated using the Tennessee Eastman benchmark process. By employing hierarchical input selection with efficient time delays, the method achieves enhanced prediction accuracy for industrial process identification.
Study of Multiple Model Predictive Control on a pH Neutralization Plant
Nonlinear processes, with their complex dynamics and sensitivity to disturbances, demand predictive models beyond linear approximations. Nonlinear Model Predictive Control (NMPC) frameworks, particularly those leveraging multiple-model and adaptive supervisory strategies, offer enhanced regulation and robustness. Recent methods based on prediction error and fuzzy weighting demonstrate superior performance in both set-point tracking and disturbance rejection compared to conventional predictive control schemes.
Multivariable input-output linearization sliding mode control of DFIG based wind energy conversion system
This work presents a control approach for maximizing active and reactive power in DFIG-based wind turbines using a multivariable input–output linearization sliding-mode strategy. By treating stochastic wind speed and aerodynamic torque as disturbances, the controller adaptively predicts and tracks maximum power operating points online. The results demonstrate that such adaptive predictive behavior ensures robust and superior performance under varying wind conditions.