1 Introduction
The soccer robot system is a highly active high-tech competitive system in recent years [1]. The system has become a standard test platform for various fields such as artificial intelligence and robotics [2, 3]. Various advanced methods and theories are used and validated [4, 5, 6]. To guide the decision system programming of mirosot, we propose a hybrid decision structure, as shown in Figure 1. The structure divides the soccer robot decision system into three levels: task determination and solution, role assignment and coordination, and robot action. The information preprocessing module is used to process the original information of the sensor, and a quick response module is used to handle the special case. The decision-making system of the Northeastern University robot soccer team is a multi-robot confrontation decision-making system based on the hybrid decision structure. Mirosot is divided into three types: small league, middle league, and large league. We use the mirosot (middle league) competition program as an example to introduce the main contents of each layer of the decision system. The project stipulates that there are 5 robots in each side of the enemy to participate in the competition.
2 Decision system description
The input of the soccer robot decision system is visual sensor information, and the wheel speed information of the robot on the field is output. In fact, the entire decision system is a mapping from sensor space to wheel speed space, defined as:
d:s-》v;
d: indicates the decision process.
s={(homerobots,enemyrobots,
Ball)}, is the sensor space containing our robot pose information, the opponent's pose information and ball position information;
Homerobots=(homerobot1, homerobot2, homerobot3, homerobot4, homerobot5).
Homeroboti (i is the robot number) is a robot structure defined as: (x, y, θ). x, y, and θ refer to the x coordinate, y coordinate, and frontal orientation of the robot, respectively. Ball is a ball structure defined as (x, y), and x and y refer to the x and y coordinates of the ball, respectively.
v={(rbv1,rbv2,rbv3,rbv4,rbv5)}, which is the robot wheel speed space. Where rbvi is our robot motion instruction with the number i. Rbvi is a velocity structure defined as (lv, rv). Lv and rv correspond to the robot's left and right wheel speeds, respectively.
The following discussion is based on two basic assumptions:
Hypothesis 1: Natural Coordinates Hypothesis: Assume that the coordinates we use in the decision system are natural coordinates;
Hypothesis 2: Right Attack Assumption: Suppose we always attack from the left half to the right half. If the actual situation is that we are in the right half, we will convert our side to the left half through a set of formulas, which are:
Object.x=bound_right-object.x; (1)
Object.y=object.y; (2)
Object.θ=π object.θ. (3)
Among them, object represents the venue object, including our robot, the opponent robot and the ball, wherein the ball structure does not contain θ, so the ball is not converted on the θ component.
Bound_right is the length in the x direction of the playing field. In the mirosot (middle league) competition, the size is 150 cm.
2.1 Robot action layer design
The action layer defines a wheeled robot action that requires multiple control cycles to complete. The action layer mainly solves the problem of robot motion control, including trajectory planning, trajectory tracking and point stabilization control.
According to the degree of complexity, we divide the action of the soccer robot into two levels: basic action and technical action. The basic action is the basis for the robot to achieve complex actions. We define three basic actions: toposition, turn, and move. The toposition is mainly used to move the robot to a given point, the turn is the robot rotating in place, and the move is the robot moving in a certain direction. Technical actions are basically built on top of basic actions. Commonly used technical actions include: goalkeeper, shoot, boundprocess, passball, blockman, pointwaiting, etc. Their functions are: goalkeeper, shot, border processing, passing, manning and running. These technical actions constitute the set of optional actions of the soccer robot, that is, the action space:
a={a|a=goalkeeper,shoot,boundprocess,passball,blockman,pointwaiting,. . .}
The design of the soccer robot's action layer belongs to the category of robot kinematics and dynamics. This layer abstracts the action of the soccer robot, encapsulates the physical model, kinematic model and dynamic model of the soccer robot, and completes the space from the robot joint (ie the wheel speed). The mapping of space to the motion space frees us from the minutiae of the robot joint space, and puts more energy into the study of the characteristics and implementation methods of a small number of robot movements.
About Silicone Camera Case
As we konw,the people more and more like take picture and the camera is hot selling in the market.In the same time,the Camera Protector Case are invented.In the camera case market,it has soft case and the hard case.Our factory is wholesales silicone Soft Camera Case .Silicone camera case has many advantages,it is not only can prevent nails from being scratched and worn for a long time, and can protect the screen and the buttons,but also has anti-skid effect,the silicone case texture is soft and comfortable,this will make you take a better picture!
Silicone Camera Case introduction:
1.Product name:Soft camera case,silicone camera case,camera protector case,Small Camera Case,Silicone Camera Cover, Camera Sleeve
2.Place of origin:Guangdong China
3.Color:any pantone color
4.Logo:Printing,debossed,embossed
5.MOQ:500pcs.
6.Package:1 pcs/opp,customized design is available.
7.Design:Customized/stock
8.Certification:FDA,LFGB,SGS,ROHS,etc.
9.Usage:Protect camera/Decoration.
10.Silicone Camera Case photo's for reference.
Silicone Camera Case
Soft Camera Case,Silicone Camera Case,Camera Protector Case,Small Camera Case,Silicone Camera Cover,Camera Sleeve
OK Silicone Gift Co., Ltd. , https://www.oemsiliconegift.com