High Effort Research Articles

Simulators and other tools in orthopedic-trauma surgery training

The classic paradigm of "learning on the patient in the operating room" is more and more in conflict with the growing requirements of cost-efficient work and patient safety. With the technology available today for simulator systems, the accessibility of digital tools and the development of ametaverse as adigital meeting place result in various application scenarios and alternatives to classic orthopedic training. First VR-desktop simulations in orthopedics and traumatology were developed more than 20years ago. VR-desktop simulators consist of acomputer with a video screen and ajoint model. Different instruments can be paired with this system and allow haptic feedback. With innovative software, numerous training programs can be selected, and the user receives precise feedback on their performance. Immersive VR simulators have also played an increasingly important role in recent years. The use of digital media such as audio and video podcasts as learning and information sources increased in the context of COVID-19. There is also an increasing number of orthopedic and trauma surgery topics on social media platforms. In all fields, however, there is arisk of the spread of misinformation. Aquality standard must be maintained. In order to evaluate simulators and their value as atraining tool, it is important to comply with various validity criteria. Transfer validity plays an essential role for clinical application. Various studies demonstrate that the skills learned on simulators can also be successfully transferred to real clinical scenarios. A lack of availability, costs and high effort are limitations of classic training methods. In contrast, there are versatile use cases of VR-based simulations that are individually adapted to the trainees and cannot endanger patients. The still high acquisition costs, technical obstacles and the not yet widespread availability are limiting factors. The metaverse still offers unimaginable possibilities today to transfer VR-based applications to experimental learning methods.

Assessment of Inspiratory Effort in Spontaneously Breathing COVID-19 ARDS Patients Undergoing Helmet CPAP: A Comparison between Esophageal, Transdiaphragmatic and Central Venous Pressure Swing.

The clinical features of COVID-19 are highly variable. It has been speculated that the progression across COVID-19 may be triggered by excessive inspiratory drive activation. The aim of the present study was to assess whether the tidal swing in central venous pressure (ΔCVP) is a reliable estimate of inspiratory effort. Thirty critically ill patients with COVID-19 ARDS underwent a PEEP trial (0-5-10 cmH2O) during helmet CPAP. Esophageal (ΔPes) and transdiaphragmatic (ΔPdi) pressure swings were measured as indices of inspiratory effort. ΔCVP was assessed via a standard venous catheter. A low and a high inspiratory effort were defined as ΔPes ≤ 10 and >15 cmH2O, respectively. During the PEEP trial, no significant changes in ΔPes (11 [6-16] vs. 11 [7-15] vs. 12 [8-16] cmH2O, p = 0.652) and in ΔCVP (12 [7-17] vs. 11.5 [7-16] vs. 11.5 [8-15] cmH2O, p = 0.918) were detected. ΔCVP was significantly associated with ΔPes (marginal R2 0.87, p < 0.001). ΔCVP recognized both low (AUC-ROC curve 0.89 [0.84-0.96]) and high inspiratory efforts (AUC-ROC curve 0.98 [0.96-1]). ΔCVP is an easily available a reliable surrogate of ΔPes and can detect a low or a high inspiratory effort. This study provides a useful bedside tool to monitor the inspiratory effort of spontaneously breathing COVID-19 patients.

Job Stress and Sleep Disturbances Among Career Firefighters in Northern California.

This study examined the association between job stress and sleep disturbance among career firefighters. A cross-sectional survey study was conducted with 154 career firefighters working in Northern California, US job stress was measured using the short form of the Effort-Reward Imbalance questionnaire and sleep was measured with the Patient-Reported Outcomes Measurement Information System Sleep Disturbance. Approximately 75% experienced sleep disturbance. For firefighters' job stress, high effort (odds ratio [OR] = 3.68; 95% confidence interval [CI]: 1.25-10.80), high effort-reward ratio (OR = 3.55; 95% CI: 1.23-10.23), and high overcommitment (OR = 9.09; 95% CI: 2.30-35.85) were significantly associated with increased likelihood of sleep disturbance, after adjustment for other factors. Job stress significantly affected firefighters' sleep health, suggesting the need to design effective health promotion interventions to reduce job stress and improve sleep quality for these public service workers.

Leafhopper males compensate for unclear directional cues in vibration-mediated mate localization

Ambient noise and transmission properties of the substrate pose challenges in vibrational signal-mediated mating behavior of arthropods, because vibrational signal production is energetically demanding. We explored implications of these challenges in the leafhopper Aphrodes makarovi (Insecta: Hemiptera: Cicadellidae) by exposing males to various kinds of vibrational noise on a natural substrate and challenging them to find the source of the female playback. Contrary to expectations, males exposed to noise were at least as efficient as control males on account of similar searching success with less signaling effort, while playing back male–female duets allowed the males to switch to satellite behavior and locate the target without signaling, as expected. We found altered mitochondrial structure in males with high signaling effort that likely indicate early damaging processes at the cellular level in tymbal muscle, but no relation between biochemical markers of oxidative stress and signaling effort. Analysis of signal transmission revealed ambiguous amplitude gradients, which might explain relatively low searching success, but it also indicates the existence of behavioral adaptations to complex vibrational environments. We conclude that the observed searching tactic, emphasizing speed rather than thorough evaluation of directional cues, may compensate for unclear stimuli when the target is near.

Multi-Domain Active Learning: Literature Review and Comparative Study

Multi-domain learning (MDL) refers to learning a set of models simultaneously, where each model is specialized to perform a task in a particular domain. Generally, a high labeling effort is required in MDL, as data needs to be labeled by human experts for every domain. Active learning (AL) can be utilized in MDL to reduce the labeling effort by only using the most informative data. The resultant paradigm is termed multi-domain active learning (MDAL). In this work, we provide an exhaustive literature review for MDAL on the relevant fields, including AL, cross-domain information sharing schemes, and cross-domain instance evaluation approaches. It is found that the few studies which have been directly conducted on MDAL cannot serve as off-the-shelf solutions on more general MDAL tasks. To fill this gap, we construct a pipeline of MDAL and present a comprehensive comparative study of thirty different algorithms, which are established by combining six representative MDL models and five commonly used AL strategies. We evaluate the algorithms on six datasets involving textual and visual classification tasks. In most cases, AL brings notable improvements to MDL, and the naive BvSB (best vs. second best) Uncertainty strategy can perform competitively with the state-of-the-art AL strategies. Besides, BvSB with the MAN (multinomial adversarial networks) model can consistently achieve top or above-average performance on all the datasets. Furthermore, we qualitatively analyze the behaviors of the well-performed strategies and models, shedding light on their superior performance in the comparison. Finally, we recommend using BvSB with the MAN model in the application of MDAL due to their good performance in the experiments.

Weakly Supervised Breast Lesion Detection in Dynamic Contrast-Enhanced MRI.

Currently, obtaining accurate medical annotations requires high labor and time effort, which largely limits the development of supervised learning-based tumor detection tasks. In this work, we investigated a weakly supervised learning model for detecting breast lesions in dynamic contrast-enhanced MRI (DCE-MRI) with only image-level labels. Two hundred fifty-four normal and 398 abnormal cases with pathologically confirmed lesions were retrospectively enrolled into the breast dataset, which was divided into the training set (80%), validation set (10%), and testing set (10%) at the patient level. First, the second image series S2 after the injection of a contrast agent was acquired from the 3.0-T, T1-weighted dynamic enhanced MR imaging sequences. Second, a feature pyramid network (FPN) with convolutional block attention module (CBAM) was proposed to extract multi-scale feature maps of the modified classification network VGG16. Then, initial location information was obtained from the heatmaps generated using the layer class activation mapping algorithm (Layer-CAM). Finally, the detection results of breast lesion were refined by the conditional random field (CRF). Accuracy, sensitivity, specificity, and area under the receiver operating characteristic (ROC) curve (AUC) were utilized for evaluation of image-level classification. Average precision (AP) was estimated for breast lesion localization. Delong's test was used to compare the AUCs of different models for significance. The proposed model was effective with accuracy of 95.2%, sensitivity of 91.6%, specificity of 99.2%, and AUC of 0.986. The AP for breast lesion detection was 84.1% using weakly supervised learning. Weakly supervised learning based on FPN combined with Layer-CAM facilitated automatic detection of breast lesion.

Optimizing robot grasping using soft origami zigzag gripper with artificial intelligence

The covid-19 pandemic leads to an increased need for automation in food industries. Current soft robots require cast moulding, high assembly effort, large actuators, and upfront actuation awareness due to the absence of artificial intelligence. Soft origami structures exhibit high levels of compliance. We developed a 3D print-in-place under-actuated soft origami zigzag gripper using TPU with artificial intelligence object pose estimation to grasp objects with high repeatability of success under different conditions. A self-designed 3D robot arm is used for object grasping. Grasping performance tests were conducted using different gripper designs, robot arm speed, and gripping force on objects with different masses and morphology. The soft origami zigzag gripper has better-grasping performance than the hard gripper as analyzed by paired t-test. The logistic model of the soft origami zigzag gripper’s grasping performance achieved an accuracy of 0.940 and AUC=0.911. Jetson Nano running AI CNN Resnet18, enhanced grasping performance with vision object classification achieved an accuracy of 0.923 and F1 score of 0.956. The AI CNN MobileNetV2 is used for object classification and experimental results showed it had lower accuracy as compared to the AI CNN Resnet18. The object classification using artificial intelligence optimized the robot grasping to improve automation in food processing and food handling to avoid contamination. All these can be used in the automation process in the food industry to overcome health and safety challenges.

The Design of a New 3D Print-in-place Soft Four-Legged Robots with Artificial Intelligence

Soft and flexible robots are designed to change their flexibility over a wide range to perform tasks adequately in real-world applications. Current soft robots require cast moulding, high assembly effort and large actuators. Soft origami structures exhibit high levels of compliance. In this paper, we designed a new 3D print-in-place soft four-legged robot (3DSOLR). Our soft legged robot is an endurance application adapted from the soft origami zigzag gripper. This novel and innovative design are inspired by the rigid joint Theo Jansen legged robot with highly adaptive 3D print-in-place soft origami legs capable of fluid motion and even surviving drop tests. The robot mechanism consists of four soft origami flexible legs driven by two DC motors. The 3DSOLR is lightweight and semi-autonomous using two Hall effect sensors and a wireless Bluetooth module. Being 3D print-in-place using Thermoplastic polyurethane also increases its durability while having flexibility, simplicity and safety. The robot also has a gripper inspired by the mandible of male European stag beetle (Lucanus cervus). These features make this robot suitable to be used in social robotics and rescue robotics applications. The transmitter program is implemented in Bluetooth serial communication using MIT App Inventor 2 smartphone apps and a microcontroller Arduino ATMEL is used as the main controller and code in Arduino IDE. It has artificial intelligence (AI) capability with ESP32 CAM onboard which has an object classification accuracy of 95.5% using custom Edge Impulse neural network MobileNetV1 96 x 96. This AI capability enhanced the robot’s capability in object classification for grasping.

Restoring Axonal Organelle Motility and Regeneration in Cultured FUS-ALS Motoneurons through Magnetic Field Stimulation Suggests an Alternative Therapeutic Approach.

Amyotrophic lateral sclerosis (ALS) is a devastating motoneuron disease characterized by sustained loss of neuromuscular junctions, degenerating corticospinal motoneurons and rapidly progressing muscle paralysis. Motoneurons have unique features, essentially a highly polarized, lengthy architecture of axons, posing a considerable challenge for maintaining long-range trafficking routes for organelles, cargo, mRNA and secretion with a high energy effort to serve crucial neuronal functions. Impaired intracellular pathways implicated in ALS pathology comprise RNA metabolism, cytoplasmic protein aggregation, cytoskeletal integrity for organelle trafficking and maintenance of mitochondrial morphology and function, cumulatively leading to neurodegeneration. Current drug treatments only have marginal effects on survival, thereby calling for alternative ALS therapies. Exposure to magnetic fields, e.g., transcranial magnetic stimulations (TMS) on the central nervous system (CNS), has been broadly explored over the past 20 years to investigate and improve physical and mental activities through stimulated excitability as well as neuronal plasticity. However, studies of magnetic treatments on the peripheral nervous system are still scarce. Thus, we investigated the therapeutic potential of low frequency alternating current magnetic fields on cultured spinal motoneurons derived from induced pluripotent stem cells of FUS-ALS patients and healthy persons. We report a remarkable restoration induced by magnetic stimulation on axonal trafficking of mitochondria and lysosomes and axonal regenerative sprouting after axotomy in FUS-ALS in vitro without obvious harmful effects on diseased and healthy neurons. These beneficial effects seem to derive from improved microtubule integrity. Thus, our study suggests the therapeutic potential of magnetic stimulations in ALS, which awaits further exploration and validation in future long-term in vivo studies.

Technological Trust Perceptions in Wearable Fitness Technology: A Person-Centred Approach

Technological trust is a key element impacting the success of a technology. This study focuses on fitness tracker users and their perceptions of trust towards the device. The aim of the study is to identify distinct subgroups of technological trust perceptions of fitness tracker users (n = 150) adopting a person-centred approach. Furthermore, the role of age, experience with the device and effort expectancy in trust perception subgroup membership are examined. Three distinct subgroups could be identified. The first subgroup is characterised by neutral to low data privacy perceptions and moderate perceptions towards reliability, validity, system capability and system transparency perceptions. The second subgroup is characterised by moderate to high trust perceptions, and the third group shows the highest trust perceptions in comparison to the other subgroups. Age and experience with the device were no significant predictors of trust profile membership; effort expectancy, however, was a significant predictor. Users who perceive that using the device does not require high effort and that using the tracker is easy are more likely to be classified in the high trust subgroup and in the moderate to high trust subgroup than in the low to moderate trust subgroup. Furthermore, differences between two latent profiles regarding their usage of the wearable in their physical activities were found.

Application of dynamic condensation for model order reduction in real-time hybrid simulations

This paper presents an algorithm for model order reduction in real-time hybrid simulations. The bottleneck of hybrid simulations are usually the finite element computations that must be performed in real time. The common approach to deal with this inconvenience is to involve powerful computing hardware. In the following paper opposite approach is presented—a new algorithm is proposed for reducing the model order. This allows to perform hybrid simulations more efficiently while maintaining high accuracy. The algorithm is based on dynamic condensation where the degrees of freedom are divided into masters and slaves, and system matrices are transformed in a way that only masters are kept. The transformation process is complex however, it is performed offline, therefore a relatively high computational effort is made a priori to make real-time calculations more efficient. The dynamic condensation algorithm was adapted to the requirements of hybrid simulations. The proper selection of masters is crucial for accuracy therefore, a novel approach based on evolutionary optimization is implemented. Numerical and experimental examples are provided. The examples prove that by implementing the proposed algorithms the following effects are achieved: (a) the time step of hybrid simulation can be significantly decreased when using implicit integration (increasing the accuracy of the measured dynamic behavior), (b) explicit integration can be sometimes implemented where it was previously not possible in real time, (c) smaller hardware resources can be involved (all computations in real-time hybrid simulations during the experiments were performed on a small microcontroller).

Optimizing Hill Climbing Algorithm for S-Boxes Generation

Nonlinear substitutions or S-boxes are important cryptographic primitives of modern symmetric ciphers. They are designed to complicate the plaintext-ciphertext dependency. According to modern ideas, the S-box should be bijective, have high nonlinearity and algebraic immunity, low delta uniformity, and linear redundancy. These criteria directly affect the cryptographic strength of ciphers, providing resistance to statistical, linear, algebraic, differential, and other cryptanalysis techniques. Many researchers have used various heuristic search algorithms to generate random S-boxes with high nonlinearity; however, the complexity of this task is still high. For example, the best-known algorithm to generate a random 8-bit bijective S-box with nonlinearity 104 requires high computational effort—more than 65,000 intermediate estimates or search iterations. In this article, we explore a hill-climbing algorithm and optimize the heuristic search parameters. We show that the complexity of generating S-boxes can be significantly reduced. To search for a random bijective S-box with nonlinearity 104, only about 50,000 intermediate search iterations are required. In addition, we generate cryptographically strong S-Boxes for which additional criteria are provided. We present estimates of the complexity of the search and estimates of the probabilities of generating substitutions with various cryptographic indicators. The extracted results demonstrate a significant improvement in our approach compared to the state of the art in terms of providing linear non-redundancy, nonlinearity, algebraic immunity, and delta uniformity.

Hotel AI service: Are employees still needed?

This research adopts the artificial intelligence (AI) device usage acceptance (AIDUA) model and draws on cognitive appraisal theory. It articulates a mediated relationship leading from social influence, hedonic motivation, and anthropomorphism to acceptance of or objection to AI device usage through the mediation of performance expectancy, effort expectancy, and customer emotion in the context of smart hotels. The study further draws on two-factor theory of emotion to identify the moderation of employee presence. Based on a survey design, our research adds to the literature by underscoring why human staff is still critical even in the smart service encounter, especially when customers are looking for superior performance with low effort when using AI devices. However, the presence of employees may be a hindrance in the low performance and high effort expectancies conditions. This research pinpoints a more precise mechanism of the role of employees and technology in delivering intended emotional valence.

On the contact between elasto-plastic media with self-affine fractal roughness

Modeling the elasto-plastic contact between rough interfaces may require high computational effort as real surfaces present broad roughness spectra. In this work, we propose an efficient multi-asperity model where each asperity follows Jackson and Green’s equations and both coupling and coalescence of contact spots are considered. In agreement with previous studies, the contact area A is found to rise linearly with the applied load. Moreover, under the assumption of yield stress independent of the asperity size, no differences are found with respect to the elastic contact solution for nanometric root mean square (rms) roughness amplitudes hrms. On the contrary, for micrometric hrms, the slope of the area-load relation is observed to increase when reducing the yield strength σY. We have also investigated the effect of increasing the rms roughness gradient of the surface hrms′ by adding fine-scale wavelengths to the roughness spectrum. Due to plastic deformations, the contact area A is found to be independent of the high-frequency cut-off of the roughness spectrum as it converges when increasing the number of roughness scales.

Inconsistency management in heterogeneous engineering data in intralogistics based on coupled metamodels

Abstract During the development of intralogistics systems (ILS), heterogeneous models are created, which represent discipline-specific views, e.g., control software developed by automation engineers or discrete-event simulation models created by simulation engineers. These models represent discipline-specific views on the system but contain overlapping information. Thereby, keeping the information in different development models consistent is challenging and currently requires high manual effort, which highly depends on the developers’ experience. To overcome this challenge, an approach to link heterogeneous model data and identify potential information inconsistencies within and between models automatically is proposed. The concept is evaluated with a use case containing three typical inconsistencies from five representative engineering models applied in ILS development.

Kinetic modeling of dynamically operated heterogeneously catalyzed reactions: Microkinetic model reduction and semi-mechanistic approach on the example of the CO[formula omitted] methanation

As the production of green hydrogen from fluctuating resources and the use of Power-to-X technologies become more and more important in the energy sector and the chemical process industries, the interest in dynamically operated reactors is increasing. In order to apply model-based optimization methods for process intensification the dynamic reactor behavior needs to be predicted sufficiently accurate by the underlying model equations, in which the reaction kinetics play a key role. Microkinetic elementary step models are able to describe the dynamic behavior on the catalyst surface. However, these models are laborious to derive, parameterize and require comparably high computational effort to solve. For these reasons, simpler semi-mechanistic kinetic models based on a rate determining step at steady state conditions are much more frequently applied in numerical studies. However, the underlying assumptions of these models are questionable for dynamic operation since sorption processes and possible changes in the reaction mechanisms are not considered. We present a semi-mechanistic kinetic model approach based on so called rate-affecting steps and the dynamic change in surface coverages, which is able to describe the dynamic behavior of the catalyst surface. The applicability of our approach is demonstrated for model reduction of an elementary step mechanism developed for the COx methanation and for kinetic model development with synthetic experimental data. Both methods result in a kinetic expression that can accurately reproduce the dynamic kinetic behavior predicted by the elementary step model, with a significant reduction in computational effort for our new semi-mechanistic model, making this new class of model approach attractive for model-based optimization applications.

Fishermen Interviews: A Cost-Effective Tool for Evaluating the Impact of Fisheries on Vulnerable Sea Turtles in Tunisia and Identifying Levers of Mitigation.

Fisheries bycatch is considered one of the main threats to sea turtles. To provide an updated assessment of the bycatch impact on loggerhead turtles in Tunisia, 483 interviews were conducted with fishermen in 19 ports along the Gulfs of Gabes and Hammamet. The interview questions included data on vessel and fishing gear characteristics, monthly fishing effort, and monthly sea turtle bycatch of the last year. Results revealed that sea turtle bycatch per unit effort (BPUE) was the highest for ray nets (0.73 turtles/vessel/day, 95% CI [0.02-1.15]), followed by pelagic longlines (0.6 turtles/vessel/day, 95% CI [0-4.23]) and shark nets (0.4 turtles/vessel/day, 95% CI [0-2.12]). However, due to the trawlers' high fishing effort, the cumulative impact of the trawl nets was the highest with an estimated number of 11,740 (95% CI [0-41,525.75]) turtles caught per year in Tunisia. Factors influencing the bycatch risk were gear characteristics (mesh size for nets, hook size for longlines, vertical opening for trawls), as well as season and location of operations. These factors will guide the research of mitigation solutions. The interviews with fishermen proved to be a cost-effective approach for the rapid assessment of sea turtle interactions with fisheries in the context of Tunisia.

Performance of helmet CPAP using different configurations: Turbine-driven ventilators vs Venturi devices

BackgroundTraditionally, Venturi-based flow generators have been preferred over mechanical ventilators to provide continuous positive airway pressure (CPAP) through the helmet (h-CPAP). Recently, modern turbine-driven ventilators (TDVs) showed to be safe and effective in delivering h-CPAP. We aimed to compare the pressure stability during h-CPAP delivered by Venturi devices and TDVs and assess the impact of High Efficiency Particulate Air (HEPA) filters on their performance. MethodsWe performed a bench study using an artificial lung simulator set in a restrictive respiratory condition, simulating two different levels of patient effort (high and low) with and without the interposition of the HEPA filter. We calculated the average of minimal (Pmin), maximal (Pmax) and mean (Pmean) airway pressure and the time product measured on the airway pressure curve (PTPinsp). We defined the pressure swing (Pswing) as Pmax - Pmin and pressure drop (Pdrop) as End Expiratory Pressure - Pmin. ResultsPswing across CPAP levels varied widely among all the tested devices. During “low effort”, no difference in Pswing and Pdrop was found between Venturi devices and TDVs; during high effort, Pswing (p<0.001) and Pdrop (p<0.001) were significantly higher in TDVs compared to Venturi devices, but the PTPinsp was lower (1.50 SD 0.54 vs 1.67 SD 0.55, p<0.001). HEPA filter addition almost doubled Pswing and PTPinsp (p<0.001) but left unaltered the differences among Venturi and TDVs systems in favor of the latter (p<0.001). ConclusionsTDVs performed better than Venturi systems in delivering a stable positive pressure level during h-CPAP in a bench setting.

Adolescent Bone Advantages 3Years After Resistance Training Trial.

We assessed maintenance of skeletal advantages 3years after completion of a 2-year, school-based, controlled exercise trial in adolescent girls. Middle-school girls participated in a resistance training program embedded in physical education classes. Effort groups (low-effort group [LO] and high-effort group [HI]) were identified; the control group (CON) participated in standard physical education at a separate school. Baseline and follow-up (FU) assessments at 6, 18, and 54 (FU3) months included densitometry, anthropometry, and questionnaires assessing physical maturity and nonintervention organized physical activity. Linear mixed effects models were fit to evaluate bone outcomes across all FU time points for CON versus LO/HI. Sixty-eight girls (23 CON/25 HI/20 LO) were 11.6 (0.3)years at baseline. Bone parameters did not differ at baseline, except femoral neck bone mineral density (LO < HI/CON, P < .05). Forty-seven participants provided FU3 assessment: 17 CON/16 HI/14 LO. After adjusting for height, gynecologic age, baseline bone, and organized physical activity, bone gains across all time points were greater for HI versus CON for legs bone mineral content, femoral neck bone mineral content/bone mineral density, and third lumbar vertebra bone mineral content/bone mineral density (P ≤ .05). At FU3, bone values were greater for HI versus CON at subhead, legs, femoral neck, and third lumbar vertebra (P < .03). Adolescent girls who exerted high effort in a school-based resistance training program demonstrated significant skeletal benefits 3years after program completion.

Implementation of Automatic DSP Code Generation Based on Closed-loop Algorithm for Three-level Inverters

Wind energy has the advantages of wide distribution, renewable, and non-polluting, so it is receiving more and more attention from more and more countries. As more and more wind power systems are integrated into the grid, it has an impact on the stability of the grid. To keep the power system stable, there is an urgent need for a grid simulator that can simulate various behaviors of the grid and test the reliability of the wind turbine before grid integration. Inverters, especially multilevel inverters, as the core part of the grid simulator, have been widely studied by scholars in recent years. However, compared to conventional inverters, multilevel inverters are characterized by high code development effort, great difficulty, and a long development period. In this paper, we adopt an automatic DSP code generation method with MATLAB hardware support package and give a complete system design method and development flow based on MATLAB and TMSF28335 automatic code generation. Finally, we take the closed-loop three-level MMC inverter as an example, propose an equalization algorithm suitable for automatic code generation for the capacitor-voltage balancing part, and verify the feasibility of the DSP automatic code generation in a multilevel inverter development. The feasibility of DSP automatic code generation in the development of a multilevel inverter is verified. The experimental results show that the proposed equalization algorithm with variable reference coefficient and DSP automatic code generation method can be used in the development of a multilevel inverter, which can improve development efficiency and reduce development costs.