Exploration on Trinity Competition Management Mode Driven by Double Cycle

The "student-centered" value orientation, the "results-oriented" implementation logic and the "continuous improvement" quality requirement under the framework of the Sydney Agreement provides important ideas and reference value for the organization and management mechanism of discipline competitions in the new era. Establish talent selection mechanism based on machine learning model and selecting potential outstanding students, "student-centered" construction diversity training courses, selected form more projects, to participate in competitions at all levels, form a trinity "course - projects - competition" competition training mode, explore the talent training mode "mentoring" establish a sustainable, the perfect talent training echelon. In order to promote the continuous improvement of competition management and the effective improvement of students' comprehensive ability, the competition effectiveness evaluation system is constructed to form a "double closed-loop" feedback mechanism for teachers and students. The "Trinity" discipline competition management mode driven by "double circulation" is shown in Figure 1.

Figure 1.

Trinity" discipline competition management mode driven by "double circulation"

Scientific selection of competition students is the key to achieve excellent results in discipline competition, respect the law of talent training, establish a reasonable and effective selection mechanism, and use prediction model algorithm to comprehensively analyze and predict the competition potential of students. Basic selection strategy is new first-year extensive propaganda and mobilization actively, uphold the principle of combining voluntary students and the teacher recommended, the first semester freshman year screening class education cultivates students enter the competition, education training class Open University mathematics foundation, main engineering material foundation, an introduction to robotics basic courses, such as after a semester of race courses, At the end of the semester, the training class was examined.

Delphi Method was used to solicit the opinions of 10 instructors with experience in guiding national competition awards [13]. For comprehensive analysis, College Entrance Examination Result, Competition Experience, Skill Certificate and Employment were selected Six key factors, Experience, Final Exam Result and Teacher Evaluation, were used as the characteristic indicators of students' competition potential. The student feature vector was constructed, and then the student sample dataset was established after normalization. A quantum particle swarm optimization Kernel Extreme Learning machine (QPSO KELM) model with good robustness, strong learning ability and high prediction accuracy was constructed to analyze and predict the competition potential of students in the training class [14], and the list of students entering the competition training class was determined according to the prediction results of the learning model.

According to the characteristics of student competition potential index characteristics, with good generalization performance, efficient learning performance of KELM (extreme learning machine), the fusion of good global optimization ability of QPSO (quantum particle swarm), optimization of KELM regularization coefficient of the model and the kernel function parameter, efficient and accurate student competition potential prediction. The specific prediction process is shown in Figure 2:

Figure 2.

Prediction process of students' competition potential

80 students in each group were extracted from the student database of grade 2019 to establish a sample data set for testing and analysis. 60 of each group were randomly selected as the training sample set, and the remaining 20 were used as the test sample set. The kernel Extreme learning machine (KELM) was trained with 60 training samples from each group as input feature vectors. The parameters of KELM model were optimized by quantum particle swarm optimization (QPSO) algorithm, and the optimal regularization coefficient C and kernel function parameter σ were found to determine the QPSO model. Through the prediction model, 20 test samples of the two groups were predicted respectively. The results show that the prediction results of the two groups are in good agreement with the actual situation, which verifies that the prediction accuracy of the model is high and the performance is stable. The prediction results of the competition potential are suitable for the selection of excellent students in the competition training class.

The training system of "course-project-competition" trinity linking with each other, progressing step by step, and multi-tiered is established, as shown in Figure 3. The teaching of competition courses focuses on the basic reserves of theoretical knowledge, engineering principles and programming algorithms required by various competitions. Project training relies on teachers and students' horizontal and horizontal projects and important competition items over the years, combined with the actual project key training methods to find problems, improve the ability to analyze and solve problems; All levels of discipline competition are the test platform of competition course teaching and project training.

Figure 3.

"Course - Project - competition" trinity competition training mode

Course training is an important link to focus on competition requirements, strengthen students' professional theoretical knowledge, "student-centered" training students' engineering practice ability and stimulate innovative thinking. Competition training courses according to students knowledge structure, skills, special characteristics respectively into electrical, mechanical structure and programming in the control group, geared to the needs of different groups of students combined with the key competition project carefully screening of curriculum content, grouping courses in all kinds of competitions, including foundation courses, classes, computer classes, control programming classes, classes, etc., all kinds of courses as shown in table 1.

For students participating in different competition projects, competition instructors or senior professional teachers will carry out "online and offline" hybrid project training in project-based and task-driven ways to meet students' independent and personalized learning. Depending on the longitudinal and horizontal projects of teachers and students and the important events over the years, we carry out various forms of project training, including centralized training, special training and module training. Centralized training for all kinds of competition universal problems training, including psychological quality, fighting spirit, teamwork spirit, anti-pressure ability, communication and coordination ability training, enhance students' self-confidence and tenacity, enhance the ability to solve unexpected problems in the process of competition. Thematic training aims at in-depth analysis of important competitions and key projects, such as "Challenge Cup" National College Student Science and Technology Competition, National College Student Robot Competition, etc., including competition rules, problem analysis, design requirements, ability requirements, etc. Module training summarizes all kinds of competitions and scientific research projects, and extracts core modules for intensive training, including wireless communication module, machine vision module, deep learning module, etc.

Competition actual combat from the school level, city level, provincial level, national level layer, to develop a school level, city level competition as the basis, provincial, national competition as the goal of the multi-layered competition plan. Relying on the national robot competition, the challenge cup "national college students college students competition, college students' mechanical innovation design competition of science and technology, industrial robot technology competition important games, such as organization to carry out field trials in advance, from electrical, mechanical structure and programming control group screening students configure reasonable team to participate in, through the selection competition, deficiencies can be found, shortcomings can be remedied, effective promotion can be promoted, and excellent students can be selected to participate in provincial and above competitions. Build diversified training courses, select multiple forms of projects, analyze competitions at all levels, and establish a "course training, project training, competition actual combat" step by step improvement mechanism.

According to the theoretical knowledge, practical ability and competition experience of students of different grades, a sustainable and self-perfect talent echelon training model of "mentoring" is established. Give full play to the leading role of senior students with rich experience in competition and previous outstanding graduates, and serve as course assistants for course training and project training to help junior students improve and grow quickly through words and deeds. In the university-level and city-level competitions, senior students will lead junior students to participate, and junior students will be the main force of the competition, to cultivate students' communication and coordination ability, psychological quality and craftsman spirit, so as to achieve the purpose of checking the effectiveness of course learning and project training and accumulate competition experience. In the provincial and national competitions, senior students are the main force of the competition, while junior students observe and study to cultivate students' fighting spirit, teamwork spirit, anti-pressure ability and practical and innovative ability.

The three-level echelon dynamic adjustment mode is established for contestants, reserve players and reserve players in each competition. The echelon classification facilitates the reverse design of training courses and training projects, the implementation of fine differentiation and personalized training, and the individualized teaching to help students achieve targeted ability improvement. The promotion mechanism between echelons provides room for the students of lower echelons to rise, and the mechanism of demotion or elimination will be started for the students whose comprehensive quality and ability does not meet the requirements of the competition. The dynamic mechanism of talent echelon can realize the dual role of motivation and constraint, enhance the enthusiasm and discipline of the competition team, and realize the self-promotion, self-development and continuous improvement of the competition team.

Subject contest results oriented, establish the competition after the evaluation on the performance evaluation index for students and teachers, for students in the subject to evaluate performance and awards competition process, advantages and disadvantages to clarify participating students, provide personalized guidance and special training, to ensure that the students can keep advantage, supplement the short board, so as to improve the students' practical ability to innovate. Based on the analysis of university competition evaluation experts and national competition instructors, the main influencing factors for the success of a specific discipline competition include the difficulty of winning the competition, the strength of teachers' guidance and the comprehensive ability of students [15]. The difficulty of winning a discipline competition mainly depends on the level of the competition (municipal, provincial or national), the difficulty of the competition (the strength of competitors, the proportion of winning, the completion rate, etc.) and the level of the award. The strength of teacher's guidance includes the depth of guidance (special discussion, drill, etc.), the breadth of guidance (time span of competition guidance) and the experience of guidance (teachers' experience in guiding this competition).

Analytic hierarchy process (AHP) is the combination of qualitative and quantitative systematic analysis method (16-17), this study adopts AHP to establish the discipline competition achievements evaluation system, analysis to determine the hierarchy and the index weight, to race results as the goal layer, comprehensive ability, and winning the difficulty as the criterion layer, under the guidance of teachers award level, guiding the depth, the psychological quality as index layer, Specific evaluation indicators are shown in Table 2.

Five university competition evaluation experts, 15 national competition instructors and 40 national competition winners were invited to conduct a questionnaire survey. The importance of criteria level indicators were compared pairwise, and the criterion level judgment matrix was obtained according to the importance scale of indicators, as shown in Table 3.

MATLAB was used to calculate the maximum eigenroot of the judgment matrix, and the corresponding normalized eigenvectors, namely, the weight of award difficulty, teacher guidance and comprehensive ability, were 0.1571, 0.2493 and 0.5936, respectively. The consistency index of offset rate, M is the matrix dimension, and the average random consistency index RI can be obtained by looking up the table [13], and then the consistency ratio can be calculated, indicating that the judgment matrix has ideal consistency. Using the same method, the maximum characteristic root of the judgment matrix of each indicator layer can be calculated, and the weight of each indicator layer can be obtained as shown in Table 4. All the judgment matrices meet the consistency requirements.

According to the evaluation results of the competition effectiveness evaluation index system, in-depth analysis of each evaluation index is helpful to clarify the advantages and disadvantages of students participating in the competition and clarify the focus of the instructor. According to the deficiencies of team cooperation ability, innovation quality ability, engineering practice ability and psychological quality in the process of the competition, we organized a special discussion to analyze the causes and countermeasures, improved the course training and project training, optimized the competition guidance method and competition strategy, and effectively formed a closed-loop feedback mechanism. By establishing the evaluation index system of competition results, the "double closed-loop" feedback between teachers and students is formed to realize the "double loop", and the continuous improvement of competition talent selection, competition training and echelon talent cultivation is promoted.