An approach for model-based reinforcement learning (MBRL) gets presented utilizing system identification and fuzzy reward formulation. Minimum order estimation gets applied first to determine the system order. This aids transfer function approximation of the linear time-invariant system. A Mamdani fuzzy inference mechanism defines the reward signal. The model obtained integrates into the Q-learning process to simulate experiences bypassing the environment. The proposed framework showcases sample-efficient learning without interacting with the true system. The applicability currently limits to linear systems. Extending the techniques to nonlinear, time-varying dynamics could increase applicability to real-world systems. This paper contributes an initial demonstration of MBRL using simple dynamics approximation and fuzzy rewards. Performance analysis provides evidence of learning improvement without explicitly communicating with the actual plant. A reduction in steady state error and settling time gets achieved.

As wheelchair technology evolves and embraces a more prominent role in assistive technology, the onset of intelligent control systems necessitates a comprehensive review from an engineering perspective. In this work, we analyze the development and the emerging trends in intelligent wheelchair control. A specific focus is provided on classifying and comparing model-driven and data-driven control methodologies. In this review, findings from a range of past contributions are examined, including conventional control theories, rule-based systems, and modern data-driven approaches that include supervised, unsupervised, and reinforcement learning control algorithms. The analysis indicates that while model-driven methods offer interpretability, data-driven techniques—in particular those leveraging machine learning—provide for a superior adaptability for navigating complex and dynamic environments. We further highlight key supporting systems found in sensors, actuators, and human-machine interfaces. Additionally, the important functionalities such as autonomous navigation and obstacle avoidance methods are identified. Our findings point to some future objectives that need to be addressed. For example, energy efficiency, robustness in unpredictable settings, computational requirements, and associated demands when utilizing data-driven methods. One of the highlighted fields of study in this work is the integration of reinforcement learning and sensor fusion, which may hold some promising results for future wheelchair technologies.

Safe and sample-conscious controller synthesis for nonlinear dynamics benefits from reinforcement learning that exploits an explicit plant model. A nonlinear mass–spring–damper with hardening effects and hard stops is studied, and off-plant Q-learning is enabled using two data-driven surrogates: (i) a piecewise linear model assembled from operating region transfer function estimates and blended by triangular memberships and (ii) a global nonlinear autoregressive model with exogenous input constructed from past inputs and outputs. In unit step reference tracking on the true plant, the piecewise linear route yields lower error and reduced steady-state bias (MAE = $0.03$; SSE = $3\%$) compared with the NLARX route (MAE = $0.31$; SSE = $30\%$) in the reported configuration. The improved regulation is obtained at a higher identification cost (60,000 samples versus 12,000 samples), reflecting a fidelity–knowledge–data trade-off between localized linearization and global nonlinear regression. All reported performance metrics correspond to deterministic validation runs using fixed surrogate models and trained policies and are intended to support methodological comparison rather than statistical performance characterization. These results indicate that model-based Q-learning with identified surrogates enables off-plant policy training while containing experimental risk and that performance depends on modeling choices, state discretization, and reward shaping.

Hollow heart, and other crop defects, can be devastating to farmers. Hollow heart is not a disease but a physiological disorder affected by temperature, soil moisture, plant density, and other factors. These defects can cause substantial annual losses for farmers. Currently, potatoes are shipped and inspected from producers to shipping points and markets. At these facilities, samples are inspected for defects. Detection of hollow heart consists of halving potatoes and visually inspecting for defects. The defect size is compared to USDA hollow heart classification charts for acceptance or rejection. An automatic, non-destructive system to identify hollow heart has the potential to improve quality. Two methods have been developed to collect data for such a system: acoustic signal capture and visual/vibration signal capture. Data is collected and stored for one potato at a time. The procedure includes the collection of weight, proportional size, and volume, as well as the generation of an acoustic sound signal through a drop test and a motion signal captured through a vision system. To simulate hollow heart, potatoes are cored and retested by producing a new set of data. Each potato is manually cut and inspected for true hollow heart. The generated data includes over 1000 samples, each comprising proportional volume, weight, proportional size, motion, and acoustic data. Such a dataset does not exist in the current literature and can serve for the development of machine learning algorithms to detect hollow heart nondestructively. In this paper, the data is also analyzed in terms of its statistical properties, as applied for possible feature engineering in machine learning.

Safe and sample-conscious controller synthesis for nonlinear dynamics benefits from reinforcement learning that exploits a model of the plant. A nonlinear mass–spring–damper with hardening effects and hard stops is considered. Two data-driven models are employed to enable off-plant training: a piecewise linear model assembled from operating-region linear descriptions and blended by triangular memberships, and a global nonlinear autoregressive model with exogenous input constructed from past inputs and outputs. Q-learning is performed with the model in the loop using an error-indexed discrete state space, a finite force alphabet, and a reward that balances absolute tracking error with its short-horizon decrease. When the trained agents are deployed on the true plant for reference tracking, the piecewise linear model tends to yield tighter regulation near the setpoint and reduced steady-state bias, while the nonlinear autoregressive route requires less prior structural knowledge and a simpler data-collection campaign, at the cost of larger residual error in the tested scenario. These findings indicate that model-based Q-learning with data-driven models enables off-plant policy learning while containing experimental risk. Observed performance reflects a trade-off between fidelity obtained from localized linearization and generality afforded by global nonlinear regression, as well as design choices in state discretization and reward shaping. Prospective improvements include adaptive membership shaping, richer regressors, and limited on-plant refinement to reduce model–plant mismatch.