CN112595324A - Optimal node wheel type mobile robot path planning method under optimal energy consumption - Google Patents

Abstract

A path planning method for an optimal node wheeled mobile robot under optimal energy consumption comprises the energy lost by motor efficiency, potential energy when the ground fluctuates, energy consumed by rolling friction resistance, speed change and energy consumed by turning, actual cost g (n), estimated cost h (n) and an estimation function f (n); in the searching process, after a passable node which is not put into an openlist table in 8 adjacent nodes of a current node is put into the openlist table, the passable node is put into a caculist table which is emptied after a new current node is searched each time, meanwhile, a node which is in the openlist table and has a lower g (n) value is also put into the caculist table, and a node with a minimum f (n) value is found from the caculist table to be used as a preselected node n_p(ii) a Judging whether the preselected node is in the obstacle range; if the caculist table is empty, the caculist table is put in the openlist tableAnd finding the node with the minimum f value as the current node, and improving the running speed of the algorithm. The method has the advantages of simple concept, convenient realization and high robustness, and is favorable for meeting the practical requirements of obstacle avoidance and energy consumption management of mobile robot operation.

Description

Optimal node wheel type mobile robot path planning method under optimal energy consumption

Technical Field

The invention relates to the technical field of path planning methods for wheeled mobile robots, in particular to a path planning method for a wheeled mobile robot with optimal nodes under optimal energy consumption.

Background

The wheeled mobile robot uses a storage battery as energy source, and the energy which can be provided by the wheeled mobile robot is limited, so that the research of the wheeled mobile robot path planning method under the condition of optimal energy consumption becomes a common focus subject of many industrial robot manufacturers and academic research groups.

At present, a plurality of methods for planning paths of wheeled mobile robots under the condition of optimal research energy consumption, such as ant colony algorithm, have different application ranges and solving characteristics. However, the existing path is not an ideal path, an energy-saving path cannot be always found, energy consumption caused by potential energy change is not considered, optimal energy consumption and the shortest path and multiple targets are not considered at the same time, so that the path planning method of the wheeled mobile robot under the optimal energy consumption deviates from the actual situation, and the optimal reference value of the energy consumption is reduced.

Disclosure of Invention

The invention provides a path planning method for a wheeled mobile robot with optimal nodes under energy consumption, aiming at overcoming the defects that when the path planning method for the wheeled mobile robot under the optimal energy consumption is researched in the prior art, a path is not an ideal path, an energy-saving path cannot be found always, energy consumption caused by potential energy change is not considered, and the optimal energy consumption and the shortest path multiple targets are not considered at the same time, so that an obstacle avoidance method of the mobile robot under energy consumption constraint deviates from the actual situation, the reference value of energy consumption constraint is reduced, and the like.

The technical solution adopted by the invention to specifically solve the technical problem is as follows: a path planning method for an optimal node wheeled mobile robot under optimal energy consumption comprises the steps that Ep represents energy lost by motor efficiency, Eh represents potential energy when the ground is fluctuated, Ef represents energy consumed by rolling friction resistance, Ek represents energy consumed by speed change and turning, actual cost g (n) from a starting node to a node n, estimated cost h (n) from the node n to a target node, and f (n) is equal to g (n) and + h (n) for an estimation function f (n) of the estimation node n.

The calculation model of the actual cost g (n) from the starting node to the n nodes is as follows.

Wherein, during the operation of the wheeled mobile robot, the energy consumption is reduced by the energy loss of factors including the efficiency of the motor

Potential energy E when the ground is undulating_hMg Δ h, energy consumed by rolling friction resistance E_f＝μmgcosθS_cnEnergy consumed by speed change and turning

η represents the motor efficiency; p represents motor power; s_cnRepresenting a distance from a parent node of the n nodes to the n nodes; v. of_nRepresenting the speed of the n node; v. of_n-1Representing the speed of the parent node of the n nodes; m represents the mass of the wheeled mobile robot; g represents the gravitational acceleration; Δ h represents a height difference from a current point to a next point; μ represents a friction coefficient; theta represents the included angle between the ramp and the horizontal plane; i represents the moment of inertia; ω represents the angular velocity.

The calculation model of the estimated cost h (n) from the n nodes to the target node is as follows.

Wherein S (n) represents n (x)₁，y₁，z₁) Node and target node n_goal(x₂，y₂，z₂) The estimated distance between the two is the diagonal distance. The calculation model of the actual cost s (n) is as follows.

Wherein S is_diagonal(n) and S_straight(n) the calculation model is as follows

S_diagonal(n)＝min(|x₁-x₂|，|y₁-y₂|)

S_straight(n)＝|x₁-x₂|+|y₁-y₂|

Wherein min represents taking the minimum value, S_diagonal(n) and S_straight(n) represents the distance moved along the diagonal and the manhattan distance, respectively.

In the searching process, an openlist table and a closed list table are provided, wherein the openlist table stores nodes to be searched, and the closed list table stores searched nodes. After a passable node which is not put into the openlist table in 8 adjacent nodes of the current node is put into the openlist table, the passable node is put into a caculist table which is emptied after a new current node is searched each time, meanwhile, a node which is in the openlist table and has a lower g (n) value is also put into the caculist table, and a node f (n) value which is the smallest f (n) value is found from the caculist table and is taken as a preselected node n_p. Judging whether the preselected node is in the obstacle range, if so, finding out the node n 'with the minimum f (n) value in the obstacle range'_pAnd n is_pThe smaller of the (n) values of (a) to (f) is taken as the current node; if not, preselecting node n_pIs directly madeThe current node is obtained, so that the situation of local optimization is avoided. And if the caculist table is empty, finding the node with the minimum f value in the openlist table as the current node. In this case, most of the nodes find the optimal nodes from 8 points at most, rather than finding the optimal nodes from the openlist table, thereby increasing the running speed of the algorithm.

The optimal node wheel type mobile robot path planning method under the optimal energy consumption is as follows.

The method comprises the following steps: and if the current node n is the first loop, the NC represents the loop times, the iterative calculation formula is that NC is NC +1, and NC is 1, the starting point serves as the current node n, and the step five is skipped to, otherwise, the next step is executed continuously.

Step two: and if the caculist table is an empty table, finding the node with the minimum value of f (n) closest to the target point from the openlist table as the current node n, jumping to the step five, and otherwise, continuing to execute the next step.

Step three: finding out the node with the minimum f (n) value in the caculist table, and taking the minimum node as a preselected node n_p。

Step four: judging a preselected node n_pIf the number of the nodes is within the obstacle range, finding out n 'of the nodes with the minimum f (n) value within the obstacle range'_pAnd n is_pIf less than n, the value of (f) is compared_pF is n'_pAs the current node n, otherwise n is_pAs a current node n; if not, then n_pAs the current node n.

Step four: judging a preselected node n_pIf the number of the nodes is within the obstacle range, finding out n 'of the nodes with the minimum f (n) value within the obstacle range'_pAnd n is_pComparing the values of (f), (n) and selecting n'_pAnd n_pThe node corresponding to the smallest value of f (n) is the current point n.

Step five: the current node n is put into the closed list and in the following loop the node will not be treated as the current node again.

Step six: an empty table caculist table is set.

Step seven: if the current node n is the target node, the path finding is successful, or the openlist table is already an empty table, the path finding is failed, the algorithm is ended, otherwise, the next execution is continued.

Step eight: searching 8 adjacent nodes of the point n in turn, if the point is an overthrowable point, placing the adjacent nodes into a closed list.

Step nine: if a neighbor node is not in the openlist table or the closed list table, the neighbor node is put into the openlist table, the value of f (n) of the neighbor node is calculated, and the parent node is set as n. If the value of g (n) is smaller in the openlist table, judging whether the value of g (n) is smaller, if so, replacing the original value of g (n) with the smaller value of g (n), introducing a scale factor alpha before estimating the cost value h (n), recalculating the value of f (n), setting a parent node as n, otherwise, looking up other adjacent nodes, and putting the adjacent nodes into a caculist table.

Step ten: and after all the adjacent nodes are checked, jumping back to the step two.

The method has the advantages that the optimal node wheel type mobile robot path planning method under the optimal energy consumption is adopted, and the problems that when the problem of research on an obstacle avoidance method of the mobile robot under the constraint of energy consumption is solved, the path is not an ideal path, an energy-saving path cannot be found always, energy consumption caused by potential energy change is not considered, the optimal energy consumption and the shortest path multiple targets are not considered at the same time, the obstacle avoidance method of the mobile robot under the constraint of energy consumption deviates from the actual situation, the reference value of the energy consumption constraint is reduced, and the like are solved. The method has the advantages of simple concept, convenient realization and high robustness, and is suitable for the practical requirements of obstacle avoidance and energy consumption management of the operation of the wheeled mobile robot.

Drawings

FIG. 1 is a flow chart of a path planning method for an optimal node wheeled mobile robot under optimal energy consumption according to the present invention;

fig. 2 is a simulation effect diagram of the optimal node wheeled mobile robot path planning method under the optimal energy consumption.

Detailed Description

The invention is further described with reference to the following figures and examples:

the optimal node wheeled mobile robot path planning method under the optimal energy consumption comprises the steps that Ep represents energy lost by motor efficiency, Eh represents potential energy when the ground is fluctuated, Ef represents energy consumed by rolling friction resistance, Ek represents energy consumed by speed change and turning, actual cost g (n) from a starting node to an n node, estimated cost h (n) from the n node to a target node, and f (n) is g (n) + h (n) for an estimation function f (n) of an estimation node n.

The calculation model of the actual cost g (n) from the starting node to the n nodes is as follows.

Wherein, during the operation of the wheeled mobile robot, the energy consumption is reduced by the energy loss of factors including the efficiency of the motor

Potential energy E when the ground is undulating_hMg Δ h, energy consumed by rolling friction resistance E_f＝μmgcosθS_cnEnergy consumed by speed change and turning

η represents the motor efficiency; p represents motor power; s_cnRepresenting a distance from a parent node of the n nodes to the n nodes; v. of_nRepresenting the speed of the n node; v. of_n-1Representing the speed of the parent node of the n nodes; m represents the mass of the wheeled mobile robot; g represents the gravitational acceleration; Δ h represents a height difference from a current point to a next point; μ represents a friction coefficient; theta represents the included angle between the ramp and the horizontal plane; i represents the moment of inertia; ω represents the angular velocity.

The calculation model of the estimated cost h (n) from the n nodes to the target node is as follows.

Wherein S (n) represents n (x)₁，y₁，z₁) Node and target node n_goal(x₁，y₂，z₂) The estimated distance between the two is the diagonal distance. The calculation model of the actual cost s (n) is as follows.

Wherein S is_diagonal(n) and S_straight(n) the calculation model is as follows

S_diagonal(n)＝min(|x₁-x₂|，|y₁-y₂|)

S_straight(n)＝|x₁-x₂|+|y₁-y₂|

Wherein min represents taking the minimum value, S_diagonal(n) and S_straight(n) represents the distance moved along the diagonal and the manhattan distance, respectively.

In the searching process, an openlist table and a closed list table are provided, wherein the openlist table stores nodes to be searched, and the closed list table stores searched nodes. After a passable node which is not put into the openlist table in 8 adjacent nodes of the current node is put into the openlist table, the passable node is put into a caculist table which is emptied after a new current node is searched each time, meanwhile, a node which is in the openlist table and has a lower g (n) value is also put into the caculist table, and a node f (n) value which is the smallest f (n) value is found from the caculist table and is taken as a preselected node n_p. Judging whether the preselected node is in the obstacle range, if so, finding out the node n 'with the minimum f (n) value in the obstacle range'_pAnd n is_pThe smaller of the (n) values of (a) to (f) is taken as the current node; if not, preselecting node n_pDirectly used as the current node, thereby avoiding the occurrence of the local optimal condition. And if the caculist table is empty, finding the node with the minimum f value in the openlist table as the current node. In this case, most of the nodes find the optimal nodes from 8 points at most, rather than finding the optimal nodes from the openlist table, thereby increasing the running speed of the algorithm.

The optimal node wheel type mobile robot path planning method under the optimal energy consumption is as follows.

The method comprises the following steps: and if the current node n is the first loop, the NC represents the loop times, the iterative calculation formula is that NC is NC +1, and NC is 1, the starting point serves as the current node n, and the step five is skipped to, otherwise, the next step is executed continuously.

Step two: and if the caculist table is an empty table, finding the node with the minimum value of f (n) closest to the target point from the openlist table as the current node n, jumping to the step five, and otherwise, continuing to execute the next step.

Step three: finding out the node with the minimum f (n) value in the caculist table, and taking the minimum node as a preselected node n_p。

Step four: judging a preselected node n_pIf the number of the nodes is within the obstacle range, finding out n 'of the nodes with the minimum f (n) value within the obstacle range'_pAnd n is_pIf less than n, the value of (f) is compared_pF is n'_pAs the current node n, otherwise n is_pAs a current node n; if not, then n_pAs the current node n.

Step four: judging a preselected node n_pIf the number of the nodes is within the obstacle range, finding out n 'of the nodes with the minimum f (n) value within the obstacle range'_pAnd n is_pComparing the values of (f), (n) and selecting n'_pAnd n_pThe node corresponding to the smallest value of f (n) is the current point n.

Step five: the current node n is put into the closed list and in the following loop the node will not be treated as the current node again.

Step six: an empty table caculist table is set.

Step seven: if the current node n is the target node, the path finding is successful, or the openlist table is already an empty table, the path finding is failed, the algorithm is ended, otherwise, the next execution is continued.

Step eight: searching 8 adjacent nodes of the point n in turn, if the point is an overthrowable point, placing the adjacent nodes into a closed list.

Step nine: if a neighbor node is not in the openlist table or the closed list table, the neighbor node is put into the openlist table, the value of f (n) of the neighbor node is calculated, and the parent node is set as n. If the value of g (n) is smaller in the openlist table, judging whether the value of g (n) is smaller, if so, replacing the original value of g (n) with the smaller value of g (n), introducing a scale factor alpha before estimating the cost value h (n), recalculating the value of f (n), setting a parent node as n, otherwise, looking up other adjacent nodes, and putting the adjacent nodes into a caculist table.

Step ten: and after all the adjacent nodes are checked, jumping back to the step two.

Claims (3)

1. A path planning method for a wheel type mobile robot with optimal nodes under optimal energy consumption comprises the steps that Ep represents energy lost by motor efficiency, Eh represents potential energy when the ground is fluctuated, Ef represents energy consumed by rolling friction resistance, Ek represents energy consumed by speed change and turning, actual cost g (n) from a starting node to a node n, estimated cost h (n) from the node n to a target node, and f (n) is g (n) and h (n) for an estimation function f (n) of the estimation node n; in the searching process, an openlist table and a closed list table are provided, wherein the openlist table stores nodes to be searched, and the closed list table stores searched nodes; after a passable node which is not put into the openlist table in 8 adjacent nodes of the current node is put into the openlist table, the passable node is put into a caculist table which is emptied after a new current node is searched each time, meanwhile, a node which is in the openlist table and has a lower g (n) value is also put into the caculist table, and a node f (n) value which is the smallest f (n) value is found from the caculist table and is taken as a preselected node n_p(ii) a Judging whether the preselected node is in the obstacle range, if so, finding out the node n 'with the minimum f (n) value in the obstacle range'_pAnd n is_pThe smaller of the (n) values of (a) to (f) is taken as the current node; if not, preselecting node n_pDirectly serving as the current node, thereby avoiding the occurrence of the local optimal condition; if the caculist table is empty, finding the node with the minimum f value in the openlist table as the current node; in this case, most of the nodes find the optimal nodes from 8 points at most, rather than finding the optimal nodes from the openlist table, so that the running speed of the algorithm is increased; the optimal node wheel type mobile robot path planning method under the optimal energy consumption comprises the following steps:

the method comprises the following steps: if the current node is a first loop, NC represents the loop times, the iterative calculation formula is that NC is NC +1, and NC is 1, the starting point serves as the current node n, and the step five is skipped to, otherwise, the next step is executed continuously;

step two: if the caculist table is an empty table, finding the node with the minimum value of f (n) closest to the target point from the openlist table as the current node n, jumping to the step five, and otherwise, continuing to execute downwards;

step three: finding out the node with the minimum f (n) value in the caculist table, and taking the minimum node as a preselected node n_p；

Step four: judging a preselected node n_pIf the number of the nodes is within the obstacle range, finding out n 'of the nodes with the minimum f (n) value within the obstacle range'_pAnd n is_pIf less than n, the value of (f) is compared_pF is n'_pAs the current node n, otherwise n is_pAs a current node n; if not, then n_pAs a current node n;

step four: judging a preselected node n_pIf the number of the nodes is within the obstacle range, finding out n 'of the nodes with the minimum f (n) value within the obstacle range'_pAnd n is_pComparing the values of (f), (n) and selecting n'_pAnd n_pThe node corresponding to the minimum value of f (n) is the current point n;

step five: putting the current node n into a closed list, and in the subsequent circulation, the node is not taken as the current node any more;

step six: setting an empty table caculist table;

step seven: if the current node n is a target node, the path finding is successful, or the openlist table is an empty table, the path finding is failed, the algorithm is ended, otherwise, the next execution is continued;

step eight: searching 8 adjacent nodes of the point n in sequence, and if the adjacent nodes are the non-leappable points, putting the adjacent nodes into a closed list;

step nine: if the neighbor node is not in the openlist table or the closed list table, putting the neighbor node into the openlist table, calculating the f (n) value of the neighbor node, and setting the parent node as n; if the value of g (n) is smaller in the openlist table, judging whether the value of g (n) is smaller, if so, replacing the original value of g (n) with the smaller value of g (n), introducing a scale factor alpha before estimating the cost value h (n), recalculating the value of f (n), setting a father node as n, otherwise, checking other adjacent nodes, and putting the adjacent nodes into a caculist table;

step ten: and after all the adjacent nodes are checked, jumping back to the step two.

2. The optimal node wheeled mobile robot path planning method under energy consumption optimization according to claim 1, wherein the calculation model of the actual cost g (n) from the starting node to the n nodes is as follows:

wherein, during the operation of the wheeled mobile robot, the energy consumption is reduced by the energy loss of factors including the efficiency of the motor

Potential energy E when the ground is undulating_hMg Δ h, energy consumed by rolling friction resistance E_f＝μmgcosθS_cnEnergy consumed by speed change and turning

η represents the motor efficiency; p represents motor power; s_cnRepresenting a distance from a parent node of the n nodes to the n nodes; v. of_nRepresenting the speed of the n node; v. of_n-1Representing the speed of the parent node of the n nodes; m represents the mass of the wheeled mobile robot; g represents the gravitational acceleration; Δ h represents a height difference from a current point to a next point; μ represents a friction coefficient; theta represents the included angle between the ramp and the horizontal plane; i represents the moment of inertia; ω represents the angular velocity.

3. The optimal-node wheeled mobile robot path planning method under energy consumption optimization according to claim 1, wherein the calculation model of the estimated cost h (n) from the n nodes to the target node is as follows:

wherein S (n) represents n (x)₁，y₁，z₁) Node and target node n_goal(x₂，y₂，z₂) The estimated distance between the two is the diagonal distance; the calculation model of the actual cost s (n) is as follows:

wherein S is_diagonal(n) and S_straight(n) the calculation model is as follows

S_diagonal(n)＝min(|x₁-x₂|，|y₁-y₂|)

S_straight(n)＝|x₁-x₂|+|y₁-y₂|

Wherein min represents taking the minimum value, S_diagonal(n) and S_straight(n) represents the distance moved along the diagonal and the manhattan distance, respectively.

Images