Path Planning of Mobile Robots Based on Improved Genetic Algorithm

: With the development of intelligent manufacturing, whether from the consideration of capacity, efficiency, or convenience, the requirements for mobile robots are increasing, reasonable regional path planning is one of the most critical needs, and a genetic algorithm is the best way to solve this problem, but in some complex working environments, traditional genetic algorithms will cause some problems, such as the path is not smooth, the steering angle is too large, the number of turns is large, etc. In this paper, an improved genetic algorithm is utilized to optimize the path-planning problem of mobile robots to circumvent the common issues arising from other approaches. The Improved Genetic Algorithm (IGA) has emerged as a significant advancement in the field of optimization techniques. By incorporating adaptive features, this refined approach yields enhanced performance and accuracy when compared to traditional genetic algorithms. Building on the foundational principles of evolutionary computation, the IGA employs innovative strategies, such as adaptive crossover and mutation operators, to navigate complex solution spaces effectively. It can also reduce computation time and increase efficiency by considering various considerations, such as environmental constraints and avoiding obstacle.

methods such as uniform crossover, where each bit of the offspring is randomly selected from the corresponding bits of the parents. Another method uses multi-parental crossover, where more than two parents are involved in the breeding process. This can increase the diversity of the population and prevent premature convergence (Xie et al., 2023). Mutation strategies in improved GA can also be more sophisticated than in Traditional GA. They include methods such as adaptive mutation, where the mutation rate changes over time based on the population's fitness. Another method is a self-adaptive mutation, where the individual's fitness determines the mutation rate. When planning the mobile robot's path in the grid environment, because its working environment is divided into several small grids, the movement path of the mobile robot will also be divided into multiple segments. At this time, the path planned based on the genetic algorithm will generally have the problem of the path not being smooth, mainly because the path is composed of multiple line segments (Sun et al., 2023). From this point of view, several line segments will form several angles; the size of this angle will directly affect the length and smoothness of the path. In addition, an optimal path, that is, the shortest path must be found from the robot's current position to the target position, which is affected by the number of turns. Therefore, how to use genetic algorithms to optimize these problems is of far-reaching significance, and this paper designs new mutation operators for planning the path of mobile robots (Liu et al., 2023;Nwankwo et al., 2023). probability, if it is less than the mutation probability, use the random number method to ensure the diversity of population genes, so theoretically, the number of individuals in layer A is less than layer B, and layer B is less C layer, so after each iteration to update the population, first divide the population into three levels of A, B, and C according to the layering method, and then select individuals to copy to the next generation according to the roulette method (Luperto et al., 2023;Zhang et al., 2022).

Fitness Function Design
In genetic algorithms, the fitness function is the criterion for judging the ability of individuals in a population to survive, and the objective function determines it. The fitness function is non-negative and does not correspond exactly to the objective function. The larger the fitness function's value in dealing with the problem, the better the effect. In the environment of mobile robots, the fitness function needs to include indicators such as path length and energy consumption. Because our goal is to make the path the shortest and at the same time make the least number of turns, we introduce the penalty coefficient a here, where the fitness function value and the objective function value present a negative correlation, the more turns, the larger the objective function value, the smaller the fitness value, the smaller the probability of being retained when selecting individuals, where the objective function is as follows Eq 2: (2) F is the objective function, d(pi, pi + 1) represents the distance between the gene point pi and pi + 1 to form a line segment, and a is the penalty coefficient, generally greater than 1 . M is the number of turns when all nodes on the path chromosome are connected, the path is straight when = 0, v is the moving speed of the mobile robot, and is the speed of the mobile robot when turning r is the turning radius.
1. If the raster sequence number of the gene location of the mutation point and the grid sequence number of the previous gene point location and the grid sequence number of the next gene point location meet one of the following two formulas:

Result and Discussion
Take a model based on traditional genetic algorithms and select the same population size and number of iterations, for example, we take the crossover probability as 0.8 and the mutation probability as 1, and the results are shown in Table 1 after MATLAB simulation. The results show a comparison of the path length, number of turns and number of three angles for the same number of iterations in the three environments From the results, it can be seen that when in an environment with three different iterations, if the path generated by the improved genetic algorithm described above is adopted, both the number of angles in the path, the length of the path and the number of turns is significantly optimized.

Conclusions
The above method of improving the mutation operator has achieved a series of optimization effects by selecting, crossing, mutating and other operations on the population individuals.