Greenhouse climate control aims to improve crop productivity while reducing resource use such as heating, ventilation, $CO_2$ injection, and supplemental lighting. Model predictive control provides a systematic framework for optimizing these inputs, but its performance depends strongly on the crop model used inside the controller. Existing greenhouse crop models, such as the van Henten lettuce model, are suitable for optimization but generally do not include variables that use fluorescence parameters and therefore, cannot be used with a crop fluorescence sensor.

This thesis investigates how fluorescence-derived crop feedback can be incorporated into a greenhouse optimal control framework. A proposed assimilation model is developed using electron transport rate (ETR) and PSII redox state $q_L$, two variables obtained from chlorophyll fluorescence measurements. The resulting Differential algebraic equations are formulated so that it can be evaluated within a gradient-based optimization framework.

A sensitivity analysis is performed to determine how variations in environmental conditions and uncertain physiological parameters affect the predicted assimilation rate. The analysis shows that the influence of the model inputs and parameters depends on the prevailing environmental and physiological conditions.

The proposed assimilation model is subsequently incorporated into the van Henten lettuce growth model and ultimately used within an adaptive economic model predictive controller. The controller determines the greenhouse inputs that maximize crop production while minimizing resource consumption. Closed-loop simulations are performed under parameter mismatch, online parameter adaptation, and direct physiological-feedback configurations. Results of the simulations show that the new photosynthesis formulation can be used in a optimal control setting and that the adaptive formulation reduces the effect of parameters mismatch ... Read more

Microgrids combining electrical and seasonal thermal energy storage offer significant operational flexibility, but their management is complicated by a severe separation of time scales: electrical components evolve over minutes, while aquifer thermal energy storage (ATES) systems charge and discharge over seasonal cycles spanning weeks to months. Conventional Economic Model Predictive Control (EMPC) cannot account for this long-term seasonal objective within a computationally tractable prediction horizon, and existing approaches close this gap through heuristic terminal costs or surrogate-derived terminal constraints, both of which require careful tuning and offer limited theoretical guarantees. This thesis investigates Policy-Guided Model Predictive Control (PG-MPC) as an alternative, in which a guiding policy embedded within the short-horizon MPC problem encodes the long-term seasonal objective, rather than relying on a precomputed reference trajectory. A combined heat and power (CHP) microgrid model, incorporating a heat pump, auxiliary boiler and chiller, battery storage, and an ATES system, is formulated using the Mixed-Logical Dynamical (MLD) framework to capture the system's discrete operational modes and nonlinear heat pump characteristics. On this shared model, a baseline EMPC controller using surrogate-derived terminal ingredients is compared against a PG-MPC controller using a simple sinusoidal dead-band guiding policy. Both controllers are benchmarked in simulation under imperfect disturbance forecasts across four scenarios spanning a range of climatological and electricity-pricing conditions, and both satisfy the end-of-year thermal balance constraint in every case. Operational cost performance is strongly scenario-dependent: PG-MPC achieves cost reductions under high price volatility and above average cooling demand, and performance comparable economically but slightly worse in terms of comfort constraint satisfaction across most scenarios. The improvement offered by PG-MPC is therefore best understood as a gain in adaptivity and robustness to realised weather and price conditions rather than a uniform cost reduction: it removes the dependence on a precomputed annual trajectory while preserving guaranteed recursive feasibility, and remains competitive with EMPC even with a deliberately simple guiding policy. The framework thus offers a tractable, theoretically grounded route to embedding seasonal storage objectives in receding-horizon microgrid control, with realised performance governed by the quality of the guiding policy and the calibration of the terminal ingredients.
... Read more

This dissertation contributes to the intersection of biological and ecological modelling on the one hand, and systems and control theory on the other. It examines domain-specific problems from the biomedical and agricultural sectors through a dynamic systems and control perspective. In each case study, we develop or adopt a mathematical model to describe the system dynamics and incorporate suitable control inputs. After analysing these models to establish key properties such as positivity, boundedness, and realistic asymptotic behaviour, we formulate practical management problems ranging from antiviral treatment planning to agricultural scheduling as optimal control problems. These formulations jointly capture economic objectives and the biological or ecological responses to inputs.

As a first contribution, we develop a novel differential equation model of mpox virus infection that captures the distinct effects of two candidate antiviral drugs. We analyse the model to understand its behaviour and apply optimal control to design multi-drug treatment strategies that minimise both viral load and drug use.

Next, in the first agricultural case study, we use a discrete-time annual weed population model with crop sowing densities as control inputs to investigate sustainable and economically optimal weed management. We establish global stability of periodic state trajectories and formulate the problem as a periodic optimal control problem, yielding sustainable solutions that maximise long-run economic profit. We then extend this principle to a broader class of systems, including weeds, disease, and soil nutrient dynamics. We introduce a transient optimisation framework based on finite-horizon optimal control with a terminal value function derived from periodic reference solutions and their convergence properties. This terminal value function enables the resulting solutions to achieve infinite-horizon performance that matches or exceeds that of periodic and myopic approaches, making them both sustainable and responsive to initial conditions. We demonstrate the methodology using various agricultural models and optimisation problems from the literature, illustrating its implementation and confirming the theoretical guarantees.

Finally, we address within-season optimal crop scheduling. We model crop growth in an intercropping system using a competitive Lotka-Volterra framework and estimate parameters from published time series data of oat and lupin in isolation, monocultures, and bicultures. We incorporate sowing and harvesting as impulsive control actions that reset biomasses and plant densities, and discretise this impulsive control system for optimal control. Unlike conventional steady-state approaches, this framework enables the optimisation of relay intercropping and produces substantially higher economic yields.

Collectively, these contributions demonstrate how systems and control theory can provide new analytical insights, theoretical guarantees, and flexible optimisation strategies for biological and ecological management, offering advanced solutions that go beyond conventional steady-state and myopic approaches. ... Read more

Greenhouses play a critical role in the world's food production by enabling controlled crop growth in diverse and often suboptimal climates. Effective climate control is essential to maximise plant growth with minimal energy and resource usage. Natural ventilation, regulated through mechanical vents, is a key component of this control. To achieve consistent and efficient control, automation was introduced. Initially, greenhouse automation began with integrating RBC into climate computers. RBC provided a simple and structured approach to greenhouse climate control, making it easy for growers to operate the climate control system. In current practice, RBC remains widely used for the greenhouse natural ventilation control. However, over time, the accumulation of additional rules and settings has resulted in complex systems that are increasingly difficult to manage. The high number of settings and rules led to inconsistent and inefficient use by growers.

This thesis proposes a NMPC approach for greenhouse ventilation control as an alternative to the complex rules from the RBC structure. The study focuses on maintaining near-optimal greenhouse climate conditions while minimising mechanical wear and tear and reducing operational control complexity. Moreover, the thesis provides a detailed analysis of the RBC rules structure and settings of the MultiMa, a commercial climate control computer.

A simplified nonlinear greenhouse model is used to estimate the greenhouse dynamics. The model parameters are estimated with seasonal weighted offline NLLS using historical data of a commercial greenhouse. The seasonal parameter estimation creates a summer and winter model. The summer model provides an accurate estimation of the greenhouse dynamics. The modified winter model has less accurate performance, but still follows the overall trends of the measured temperature and humidity.

The models are used to simulate various NMPC approaches and compare them to the RBC system. The performance of the simulations is evaluated based on reference tracking accuracy and the total number of vent position changes, representing the ability to maintain climate conditions and minimise the wear and tear of the vents. The results of the simulation show that NMPC can achieve similar climate conditions to measured RBC climate conditions taken as reference, while significantly reducing the wear and tear. Moreover, NMPC can successfully integrate dynamic constraints, based on the current RBC constraint rules, into the optimisation.

Overall, the proposed NMPC framework presents promising results as an alternative to the complex rules and settings of the traditional control system. The \ac{NMPC} approach offers a more manageable and interpretable control system, while maintaining desired climate conditions in the greenhouse and minimising mechanical wear and tear. ... Read more

This thesis aims to develop an optimal control strategy for the DOT 500kW Pilot Reverse Osmosis (DOT500PRO) turbine system. The system integrates a 500 kW wind turbine with a reverse osmosis (RO) module to produce freshwater. The primary goal is to maximise revenue generation by optimising the turbine’s state transitions based on wind predictions.
The thesis begins with an analysis of the DOT500PRO and its state machine, identifying operational states, transitions, and constraints. A Markov model is used to model and predict wind speeds, which fits nicely with the Markov Decision Process (MDP) framework. The problem is formulated as an MDP, and multiple control strategies, including Threshold Control, Model Predictive Control (MPC), Stochastic Dynamic Programming (SDP), and Approximate Dynamic Programming (ADP), are evaluated.
MPC is found to be computationally intensive, making it less feasible for real-time control. SDP shows promising results, but is limited by the curse of dimensionality, restricting the use to higher order models. ADP, which approximates SDP solutions, can offer a potential controller for higher order models but requires further tuning and optimisation.
Simulations are conducted to compare the performance of these control strategies in several scenarios. While SDP demonstrates slight improvements over threshold control on the training dataset, its performance on different wind patterns is less consistent. The study concludes that while proactive control strategies such as SDP and ADP can offer improvements over reactive methods, their performance is dependent on the accuracy of wind predictions and the specific operational conditions.
Future work suggestions include refining the turbine and wind models, exploring adaptive control methods, and conducting real-life experiments to validate the control strategies, which are crucial for practical implementation and optimisation. ... Read more

Congestion on the electricity grid is a growing issue in the Netherlands. To reduce stress on the grid, electricity providers are introducing new pricing structures. This thesis considers a pricing model that combines day ahead electricity prices with an added monthly peak demand charge. The challenge is that classical heating, ventilation, and air-conditioning (HVAC) control strategies are not responsive to these pricing mechanisms, leading to increased operational costs.
This project focuses on a heat pump-based floor heating HVAC system designed to maintain thermal comfort in a small office environment. Economic model predictive control (EMPC) can leverage the building’s thermal mass to exploit variations in time-of-use (TOU) pricing. However, the monthly peak demand charge complicates the control problem, as its long-term effect is difficult to capture with the relatively short prediction horizons typical of MPC. To address this limited horizon, a method is proposed that artificially extends the prediction horizon to better account for long-term cost associated with the peak demand charge.
This artificially extended horizon is enabled by leveraging historical data for both the disturbance and electricity prices, allowing it to be approximated offline as a terminal cost function for the MPC. The proposed method, incorporating this artificial horizon extension into the MPC formulation, is compared to alternative strategies that address the peak demand charge by scaling it down within the cost function based on the length of the prediction horizon.
All methods are evaluated using information on electricity prices and disturbances that would realistically be available in practice. Results show that the MPC approach incorporating a terminal cost function economically outperforms methods that apply aggressive down scaling of the peak demand charge. However, similar economic performance is observed when compared to a more conservative scaling approach. Highlighting how conservatism with respect the peak demand is beneficial when operating under uncertainty
... Read more

Several stochastic model predictive control schemes are formulated to reduce constraint violations in the economic control of the climate in a lettuce greenhouse under weather forecast and parameter uncertainty. The schemes are tested in simulation. Two separate approaches are taken in the formulations. The first involves analytical constraint tightening through system linearization. Linearizing the system around the trajectory is found to improve performance compared to linearizing around a point. The linearized schemes proved to be overly conservative, especially under parameter uncertainty. The second approach is through tracking the average constraint violations to formulate adaptive constraints which do not require prior information about the underlying uncertainties. Originally proposed for linear systems, this approach is simplified and modified to impose a constraint tightening on deterministic nonlinear model predictive control. The adaptive schemes improve constraint compliance with reduced conservatism leading to a more acceptable increase in input costs compared to the linearized schemes. The results indicate that adaptive average violation constraints may be a useful tool in stochastic model predictive control and warrant further investigation. ... Read more

Active knee prostheses are potent in assisting users, providing symmetry in walking, reducing metabolic costs, and preventing long-term health problems. The heart of their complex control algorithm employs the Impedance Control (IC) Law, which controls the torque output of the device by three parameters: stiffness coefficient, equilibrium angle, and damping coefficient. Ideally, these parameters should be personalized and adaptive to address interpersonal and intrapersonal variations on level-ground walking. However, current practices achieve personalization only through basic normalization and do not address adaptiveness. This thesis aims to utilize an RL framework with a novel frequency-domain state representation to address personalization and adaptiveness simultaneously. A policy iteration algorithm from the Q-learning family was chosen as the essence of the RL framework, and bellman error (BE) was chosen as the primary evaluation metric. The study revolves around two hypotheses. First, does the RL framework is suitable for the system? Second, does the frequency-domain state representation perform better than the time-domain state representation? Within the scope of the thesis, a custom environment was created by modifying the Humanoid-v04 environment of OpenAI Gym using MuJoCo (Multi-Joint dynamics with Contact). This environment is used to train the RL framework and conduct the experiments. Results suggest that the proposed RL framework can improve the system, and the frequency-domain state representation is superior to its time-domain counterpart. The latter conclusion has an impact beyond the active knee prosthesis domain and can inspire any trajectory following tasks with periodic signals. ... Read more

Greenhouses allow production of crops that would otherwise be impossible. Permitting more local, fresher and nutrient richer crop production. Eorts are taken to minimize societal harm due to energy and resource consumption by greenhouse production systems. One way to control such systems is by using model predictive control. Optimal crop yield and resource eciency can, in theory, be achieved by model predictive control. Unfortunately, one major drawback of model predictive control is that it is not well equipped to deal with parametric uncertainty. Significant prediction errors can occur when a mismatch between the model and the real system exists, resulting in deteriorated performance of the system. Strategies exist, such as robust MPC, that are designed to handle uncertainty, but those often result in conservative control policies. This thesis proposes to use model predictive control as a function approximator for RL in order to learn values for model and MPC parameters that can deliver optimal performance in the case of model mismatch.
In this thesis, data-driven economic nonlinear model predictive control using Q-learning is proposed as a method to alter the model parameters. The performance of the system af- ter learning is compared to approaches using robust and nominal model predictive control. Three dierent goals are determined: maximizing economic profit, minimizing the constraint violations and maximizing the economic performance while minimizing constraint violations.
In this work, an ENMPC scheme is used as a function approximator in a Q-learning envi- ronment. The optimization solution from the ENMPC scheme is used as the input to the system, while the Q-learning agent optimizes the parameter values of the ENMPC scheme and model for the environment. The performance of the system after learning is compared to approaches using robust and nominal model predictive control. The simulation results show that the data-driven ENMPC using reinforcement learning is able to decrease constraint vi- olations by up to 94%, but unable to increase economic performance compared to nominal MPC, compared to robust MPC the EPI is increased by almost 10% while keeping constraint violations at a similar level. ... Read more

Quadrupedal robots possess the ability to move freely in the world and perform a variety of actions that would be unsafe or impractical for humans to perform. In the SNOW project, a quadrupedal robot is tasked with aiding firefighters in rescue missions during house fires by locating humans and assessing their health. Pushing away obstacles that cannot be circumvented otherwise is one of the many capabilities a quadrupedal robot should possess to be of most use in such missions. We develop a proof of concept by solving two problems sequentially: which stance to take on prior to the push and how to perform the push.

The process of stance selection starts with generating a certain amount of stances. Stances are generated starting from a preselected stance appropriate to the goal location and deviating from the 12 joint angles with a normal distribution. All generated stances are ran through a number of filters, which rely on solving for the forward and inverse kinematics of the robot. These filters check if the initial position is sensible and balanced and if the projected final position is close to the goal and balanced. The inverse kinematics are solved using Adaptive-Network-Based Fuzzy Inference Systems (ANFIS), which results in accurate estimations within a time frame that can be used in real-time applications. The final stance is selected by comparing the total displacement of all joint angles per stance, where the lowest total displacement is considered optimal.

The push is controlled by a nonlinear model predictive controller. We strive for a stable push, where the contact between the pusher and the object sticks, by keeping the movement of the end-effector within the motion cone. The motion cone denotes all twists the object can have without slipping at the contact with the pusher and is constructed using the limit surface to model the interaction between the object and the support surface and the generalized friction cone to model the interaction between the pusher and the object. We find that the motion cone predicts stick and slip with an accuracy slightly higher than 80%. Our controller steers accurately to all goals that lie within the motion cone and moves the object with a twist on the edge of the motion cone if the goal location lies outside of the motion cone. The robot remains balanced throughout the pushing motion in the vast majority of cases, but is more at risk of tipping over when pushing heavier objects. ... Read more