Tuning nonlinear model predictive control via Bayesian optimization: a comparative performance analysis

One of the crucial issues in developing nations is diminishing the yield gaps. Therefore, accurate yield gap estimation has many real-world uses for increasing crop production. Utilizing comparative performance analysis (CPA) techniques, the yield gap of wheat fields was evaluated in this study. In Varamin, Tehran Province, Iran, data on 104 wheat fields were collected between 2018 and 2020 and every aspect of wheat field management has been documented. The CPA model determines the yield gap’s contributing factors and potential yield. The results of data analysis revealed that the production ranged from 2600 to 7600 kg ha−1. The CPA method predicted a potential yield of 9316 kg ha−1 and found a yield gap of 3748 kg ha−1; this amount was 40.23% of the potential yield. Leaf chlorophyll (29%), irrigation at stem extension (9%), LAI (7.7%), soil salinity (8.2%), field area (16.3%), phosphorus consumption (6%), nitrogen utilized at the stage of tillering (16%), and HI (7.8%) all contributed to the yield gap in the CPA. It has been said that the computed yield in CPA is a potential yield that can be reached. CPA is a cheap and straightforward tool that could identify yield gaps and their causes in a district without the need for costly experiments. Therefore, developing nations with significant efficiency and yield gaps can use these techniques effectively.

Narrowing the yield gaps is one of the major concerns in developing countries. Closing yield gap to obtain attainable yield is a viable option for providing information regarding the reason of yield loss. Hence, accurate estimation of the yield gap has many practical applications for enhancing production of crops. This research was conducted for assessing the yield gap of rice-grown fields using boundary-line analysis (BLA) and comparative performance analysis (CPA) methods. Collection of 100 rice-grown fields data were done in Sari region, Mazandaran province, one of the major rice production areas in northern Iran from 2015 to 2016. All paddy field management operations from preparation of nursery to harvest of yield has been recorded for local rice varieties. The CPA model calculate the potential yield and factors causing yield gap. In contrast, BLA model were fitted to the edge of data cloud of rice yield versus field managing variables from monitoring. Analysis of data in 100 monitored paddy fields demonstrated that rice yield varied from 3100 to 5430 kg ha−1. Prediction of potential yield for CPA and BLA methods were 5703 and 5369 kg ha−1, respectively. The yield gaps calculated by CPA and BLA methods in 1212 and 881 kg ha−1, respectively. In the CPA, the share of yield gap for variables entered in the model were 5% for cover crop of canola, 18% for legumes before rice cultivation, 4% for seed disinfection, 10% for seeding date in nursery, 11% for seedling age, 11% for seedling growth stage for transplanting, 5% for mechanized transplanting, 4% for fertilizer top-dressing, 27% for number of top-dressing and 6% for foliar application of nutrients. In the BLA, an average attainable yield, based on the optimum level of the 12 studied variables, was 5369 kg ha−1 with an 881 kg ha−1 yield gap. Regarding the fact that calculated yield in CPA and BLA, it has been stated that this potential yield is attainable. CPA and BLA are cheap and simple tools that, without the need for expensive experimentation, is able to detect yield gap and its causes in a district. Therefore, these methods can be used effectively in developing countries where the highest yield gaps exist.

Nonlinear optimal control for robotic exoskeletons with electropneumatic actuators

Due to their efficiency and high power output, diesel engines find extensive use in the automotive, transportation, industrial, and agricultural sectors. However, these engines encounter several challenges, including the emission of pollutants such as nitrogen oxides and particulate matter, as well as their reliance on fossil fuels. As a result, the demand for alternative fuels has risen significantly. Biodiesel, derived from various sources, has emerged as a promising substitute for diesel fuel. Among these alternatives, mango kernel biodiesel is currently being investigated as a renewable fuel option for diesel engines. In this current research study, a single-cylinder diesel engine was used to investigate the effects of mango kernel biodiesel (B10) as fuel compared to conventional diesel fuel. The engine was operated under different loading conditions (25%, 50%, 75%, and 100%) and varying fuel injection pressures (400 bar, 500 bar, and 600 bar), while maintaining a compression ratio of 18. The research focused on conducting a comparative analysis of engine performance, and emissions between the two fuels viz. conventional diesel fuel and mango kernel biodiesel blend. For major test cases, the engine recorded higher brake thermal efficiency (BTE) and lower brake specific fuel consumption (BSFC) as compared to the biodiesel blend. At full load and higher injection pressure, the B10 blend increased BTE by 4.83% and decreased BSFC by 5.40% than diesel. The smoke formation, CO, HC emissions were notably higher with B10 blend.

In this paper, the parameterized dynamical model of the diesel engine intake and exhaust system using a data-based method, namely a Gain-Scheduled model is proposed and designed based on the data from a virtual engine test bench under normal load conditions. In the first step, the Multiple Input Multiple Output model structure is defined with five inputs and two outputs. Using the constructed model, it is possible to establish the relations between intake Manifold Pressure, Air Mass Flow, the control signals, and changes of the load. Then, the model is further used to design a Nonlinear Model Predictive Control controller, aimed at optimizing the efficiency of the combustion system in terms of the control reference value tracking with respect to emission reduction. This paper follows a model-based design approach to construct the Nonlinear Model Predictive Control objective function for the engine intake Manifold Pressure and Air Mass Flow nonlinear control problem. The proposed data-based dynamical modeling method is shown to increase the flexibility for the modeling of nonlinear plant at a low cost in computational requirements. The experimental results illustrate that the optimized nonlinear control approach significantly improve the control reference tracking performance and the exhaust emissions against the standard decentralized Single Input Single Output control in the standard production Engine Control Unit.

<div>With the rapid development of autonomous driving technology, unmanned ground vehicles (UGVs) are gradually replacing humans to perform tasks such as reconnaissance, target tracking, and search in special scenarios. Omnidirectional mobility based on rapid adjustment of vehicle heading posture enhances the applicability of UGVs in specialized scenarios. Omnidirectional mobility signifies the capability for rapid adjustments to the vehicle’s heading angle, longitudinal velocity, and lateral velocity. Traditional vehicles are constrained by the limitations of under-actuation, which prevents active regulation of lateral movement. Instead, they rely on the coordinated regulation of longitudinal and yaw movements, failing to meet the requirements for omnidirectional mobility. Distributed vehicles featuring steering distributed between the front/rear axles and four-wheel independent drive leverage the over-actuation advantages provided by multi-actuator coordinated control, making them particularly suitable for omnidirectional mobility at large sideslip angles. This feature enables the UGVs to achieve rapid adjustment of vehicle heading posture. However, existing control strategies centered on stabilizing yaw rate and suppressing sideslip angles cannot adapt to the decoupling control requirements of such platforms. Additionally, the strong coupling characteristics between actuator subsystems further exacerbate control difficulties. To this end, this article proposes a full-state decoupling motion control strategy, the nonlinear model is locally linearized at each equilibrium point of the vehicle, and a set of equilibrium state models is derived. The validity of this local linearization method is verified through phase diagram analysis and modal analysis. The Bayesian optimization (BO) algorithm is then employed to optimize and identify the cornering stiffness of the front/rear axles at each equilibrium point in these locally linearized models, thereby enhancing the characterization ability of the linear model for the nonlinear dynamic model at the corresponding equilibrium points. Subsequently, a full-state decoupling motion controller is designed by integrating the model predictive control (MPC) algorithm. Finally, the controller presented in this article is employed on the distributed vehicle experiment platform (DVEP). The experimental results demonstrate that in two drift-like scenarios with different sideslip angles, compared with the baseline controller, the path tracking error of this method is reduced by more than 13%, and the sideslip angle tracking error is reduced by more than 12%.</div>

A nonlinear optimal control approach for voltage source inverter-fed three-phase PMSMs

Bayesian optimization with reference models: A case study in MPC for HVAC central plants

Prediction of outcomes in chronic lymphocytic leukemia patients treated with ibrutinib: Validation of current prognostic models and development of a simplified three-factor model.

In this paper, we discuss the performance of non-linear model predictive control (NMPC) under different learning methods.Considering the difficulty of modeling nonlinear systems, the accuracy of the predictive model greatly affects the performance of the system. In addition, online optimization algorithms also greatly affect the efficiency of NMPC. This paper combines support vector machine (SVM) with Bayesian optimization, and uses a multi-step forward support vector machine prediction model to solve the NMPC problem. Use this strategy to narrow the selection range of optimal parameters, and finally find the optimal parameters. Compared with SVM, the performance of Multi-Layer Perceptrons(MLP) is poor, but the method still performs well. In addition, after comparing many different kernel functions, we believe that due to the existence of nonlinearity, the Gaussian kernel has better performance than other kernel functions. Finally, through the simulation research of the nonlinear system, the effectiveness of the two control schemes is verified. It should be noted that this article does not propose a new solution, we just compare the performance of different learning methods.

Engineering students are challenged with implementing and developing systems within STEM disciplines. The dialectic design approach and comparative performance analysis were created for undergraduate engineering students as a teaching method to facilitate and improve student-learning experiences in STEM disciplines. We had found in our study that both the dialectic design approach and comparative performance analysis are critical to the theoretical development and the fundamental practices for engineering education in course learning objectives. These teaching methods were created for undergraduate engineering students to support specific interdisciplinary practices such as aviation sciences and course objectives focused on emerging issues concerning the design process and performance analysis. An undergraduate engineering course must promoted student-learning experiences for innovative practices through engineering models and performance analysis. The integration design in this course supported areas that include complex aviation science projects and the requirement constraints for system development.

To reduce the yield gap, specifying yield constraints in a particular area is necessary. A complete yield gap assessment method must provide information regarding potential yield, actual yield, and causes of the gap and their importance. Therefore, documenting the production process to explain crop management factors in each area is very important. The objective of the study was to perform a rice yield gap analysis by using comparative performance analysis (CPA) and boundary-line analysis (BLA). Data were gathered from about 100 paddy fields in Neka, eastern Mazandaran province, one of the major rice producing regions in Iran, in 2015 and 2016. All agricultural practices from nursery preparation to harvest have been recorded for improved rice cultivars. CPA focuses on the ability to estimate potential yield and the reason for a yield gap. Boundary lines were fitted to the edge of the data cloud of crop yield versus management variables in data from paddy fields monitoring. The documenting analysis shows that the range of paddy yield in 100 fields varied from 6100 to 8200 kg ha–1. Potential yields were 9241 kg ha–1 for CPA method, and 7999 kg ha–1 for BLA method. Furthermore, yield gap predicted 2047 kg ha–1 for CPA method and 874 kg ha–1 for BLA method. In BLA, the average relative yield and relative yield gap of the 13 investigated variables were 89.75% and 10.25% respectively. These results show the importance of each management factor in yield gap. It was concluded that CPA and BLA as applied in the study is a cheap and simple method that, without the need for expensive experimentation, is able to detect yield gap and its causes in a district. From these results, it can be said that the calculated yield gap is close to the definition given for the utilised yield gap and shows the difference between the actual yield and attainable yield in relation to the environmental conditions of the region.

Detailed and reliable information on land use systems, as needed for quantitative studies, is scarce and often of low quality. This calls for (guidelines on) data harmonization. Practical concepts to describe and study land use are discussed; the development of the Land Use Database software was instrumental in defining them. Required is that by plot, information on land use purpose(s), on operations and on observations as made by land users is put on record through interviews. To classify land use, three types of classifiers are available: purpose, operation sequence, and context classifiers; using them keeps the possibility to prepare user-defined classification systems open.<br/><br/>Detailed land use descriptions augment Land Use Type (LUT) concepts presented by the FAO Guidelines for Land Evaluation. Biophysical LUT requirements emphasized in land evaluation studies are often crop requirements with management requirements predominantly of a socio-economic nature. The FAO guidelines make insufficient use of information on land use operations that are applied to overcome land aspects that limit yields or reduce production. Proposed is a procedure to evaluate practical technology options to remedy limiting conditions.<br/><br/>Quantitative production functions are not standard output of land evaluation studies. Use of simulation models for quantitative studies is restricted because presently they can not capture the full dynamics of yield limiting and yield reducing factors and can not consider all management options.<br/><br/>Many actual production situations face yield constraints that cause a considerable gap between actual yields and yield levels possible with improved technology. Yield gap studies are essential to identify the biophysical factors and cultural practices that cause the gap. Comparative Performance Analysis (CPA) is an approach to study yield-gaps; it defines quantified yield-gap functions. The key feature of CPA is to relate, after surveying on-farm production situations, differences in land and land use to differences in system performance. CPA complements established land use study methods. The Land Use Database supports it.<br/><br/>Three CPA studies are included. The CPA study on rice identifies priority areas for development. It explains 83% of the yield variability across 63 sites. The CPA study on mango was undertaken to remedy the "trial and error" type of management practised in the study area. It identifies the relative importance of selected production factors, i.e. soil-related (30%), management-related (49%), and crop-related (21%). The CPA study on the impacts of land use on the environment evaluated the merits of four erosion indicators. The indicators could function as Land Quality Indicators to reflect soil loss over time. Pre-rills are promising as an indicator. Their occurrence gave the best correlation to management related site conditions. The relation prepared was not map unit specific and suggested that combined positive conditions reduce the formation of pre-rills exponentially.

Ride comfort is an emerging focus in autonomous driving, yet integrating passenger behavior into vehicle motion planning remains challenging due to the impracticality of real-time passenger state feedback and high computational demands. This paper proposes a solution that approximates Ideal Motion Planning (IMP)—which traditionally requires nonlinear model predictive control (NMPC) and passenger state feedback—using linear model predictive control (LMPC) with weight tuning via Bayesian optimization. Our method eliminates the need for passenger state feedback and reduces the computation time, allowing real-time implementation. Although simplifying the cost function to prioritize tracking and stability, we achieve vehicle motion that maintains ride comfort equivalent to IMP, as demonstrated in simulations. This approach offers a practical pathway to improve passenger comfort in autonomous vehicles without additional sensory input.

This article presents the design, simulation and real-time implementation of a constrained non-linear model predictive controller for a coupled tank system. A novel wavelet-based function neural network model and a genetic algorithm online non-linear real-time optimisation approach were used in the non-linear model predictive controller strategy. A coupled tank system, which resembles operations in many chemical processes, is complex and has inherent non-linearity, and hence, controlling such system is a challenging task. Particularly important is low-level control where often instability and oscillatory responses are observed. This article designs a wavelet neural network with high predicting precision and time–frequency localisation characteristics for an online prediction model in the non-linear model predictive controller to show the effectiveness of this approach in controlling the liquid at low level. To speed up the training process, a fast global search stochastic non-linear conjugate wavelet gradient algorithm is initially used to train the wavelet neural network structure before the genetic algorithm optimisation technique is utilised to tune adaptively the wavelet neural network parameters. The non-linear model predictive controller algorithm is tested for both approaches: first, in a simulation using identified models, and second, in a real-time practical application to a single-input single-output system coupled tank system. The results show an excellent control performance with respect to mean square error and average control energy values obtained.