Search

CN-121995925-A - Intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion

CN121995925ACN 121995925 ACN121995925 ACN 121995925ACN-121995925-A

Abstract

The invention discloses an intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion, and belongs to the technical field of robot safety control and obstacle avoidance. The method comprises the steps of obtaining environment sensing information and odometer information, generating a target list and a local environment occupation grid, obtaining gesture state information of an actuating mechanism, determining external envelope parameters such as an overhanging length, an overhanging direction angle and the like, expanding the local environment occupation grid to obtain the gesture envelope occupation grid according to the external envelope parameters, carrying out self-adaptive adjustment on a classification security domain threshold according to the external envelope parameters, judging an early warning domain, a deceleration domain and a parking domain, and outputting a mobile chassis speed/braking control instruction and an actuating mechanism locking/recycling linkage instruction. The invention can dynamically adjust the safety domain along with the gesture change of the executing mechanism, reduce false triggering and unnecessary parking, and improve the obstacle avoidance safety and the operation efficiency in a complex environment.

Inventors

  • HUANG HE

Assignees

  • 深圳市小楠智能制造有限公司

Dates

Publication Date
20260508
Application Date
20260212

Claims (9)

  1. 1. Intelligent sorting robot developments obstacle avoidance system based on multisensor fuses, sorting robot is used for carrying out pushing away branch/water conservancy diversion operation to the parcel, its characterized in that includes: The perception fusion unit is used for generating a target list of dynamic obstacles based on the environmental information of at least one visual sensor and at least one distance sensor and combining with the odometer information of the mobile chassis, and generating a local environment occupation grid; the gesture envelope parameter unit is used for acquiring gesture state information of the push-off/flow-guiding executing mechanism and determining outline envelope parameters according to the gesture state information; The gesture association safety domain unit is used for performing obstacle expansion processing on the local environment occupation grid according to the profile envelope parameters to obtain a gesture envelope occupation grid, adjusting a threshold value of the gesture association safety domain based on the profile envelope parameters, and judging an early warning domain, a deceleration domain and a parking domain according to the gesture envelope occupation grid and the adjusted threshold value; the linkage control unit is used for outputting a speed limiting instruction and/or a braking instruction of the mobile chassis according to the hierarchical security domain and outputting a locking instruction and/or a recycling instruction of the push-split/flow-guiding executing mechanism; The hierarchical security domain threshold tightens as the profile envelope parameter increases, and an amount of inflation of the obstacle inflation process is determined by the profile envelope parameter such that the hierarchical security domain decision is simultaneously constrained by the pose envelope occupancy grid and the pose association threshold.
  2. 2. The intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion according to claim 1, wherein the profile envelope parameters at least comprise an overhanging length and an overhanging direction angle of a push-off/flow-guiding executing mechanism, and optionally comprise overhanging side and profile envelope expansion, the gesture state information comprises encoder readings and/or limit switch states of the executing mechanism, and the profile envelope parameters are obtained by converting encoder readings and/or selecting a preset parameter table through limit switch states.
  3. 3. The intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion according to claim 1, wherein the obstacle expansion process is a morphological expansion of the occupied grid, the adopted structural element is a directional ellipse, the long axis direction of the structural element is consistent with the overhanging direction angle, the expansion amount in the overhanging direction is larger than the expansion amount in the vertical overhanging direction, and the long axis length is increased as the overhanging length is increased.
  4. 4. The intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion according to claim 1, wherein the hierarchical security domain threshold comprises at least a deceleration threshold and a parking threshold, the threshold is calculated based on time-to-collision TTC and/or minimum security distance, and is satisfied that when the overhanging length increases, at least one of the deceleration threshold and the parking threshold increases monotonically, the threshold adjustment is implemented by a linear function or a piecewise linear function, and the threshold maximum and minimum are constrained by a preset security range.
  5. 5. The intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion according to claim 1, wherein the linkage control unit outputs a diversion mechanism locking instruction when judging that the intelligent sorting robot enters a deceleration domain, and the intelligent sorting robot dynamic obstacle avoidance system is executed in the following sequence when judging that the intelligent sorting robot dynamic obstacle avoidance system enters a parking domain, wherein the intelligent sorting robot dynamic obstacle avoidance system firstly outputs a moving chassis braking instruction to enable the chassis speed to be lower than a preset low-speed threshold value, then outputs a diversion mechanism recycling instruction, and allows switching from the parking domain to the deceleration domain or an early warning domain and applies a slow starting speed upper limit after the diversion mechanism recycling is completed or reaches a preset safe posture.
  6. 6. The intelligent sorting robot dynamic obstacle avoidance method based on multi-sensor fusion, wherein the sorting robot is used for performing pushing separation/diversion operation on packages, and is characterized by comprising the following steps of: S1, acquiring environment information of at least one visual sensor and at least one distance sensor, and acquiring odometer information of a mobile chassis; s2, carrying out fusion processing based on the environment information and combining the odometer information, generating a target list of dynamic obstacles, and generating a local environment occupation grid; s3, acquiring attitude state information of the push-off/flow-guiding executing mechanism, and determining appearance envelope parameters according to the attitude state information; s4, performing obstacle expansion processing on the local environment occupation grid according to the shape envelope parameters to obtain a gesture envelope occupation grid; s5, adjusting the threshold value of the classified security domain based on the profile envelope parameter, and judging an early warning domain, a deceleration domain and a parking domain according to the grid occupied by the gesture envelope and the adjusted threshold value; s6, outputting a speed limiting instruction and/or a braking instruction of the mobile chassis according to the early warning domain, the deceleration domain and the parking domain, and outputting a locking instruction and/or a recycling instruction of the push-separation/diversion executing mechanism; wherein the hierarchical security domain threshold tightens as the profile envelope parameter increases and an amount of inflation of the barrier inflation process is determined by the profile envelope parameter.
  7. 7. The method for dynamically avoiding barriers for intelligent sorting robots based on multi-sensor fusion according to claim 6, wherein step S4 comprises the steps of morphologically expanding the occupied grid by adopting directional elliptic structural elements, wherein the long axis direction of the structural elements is consistent with the overhanging direction angle, the expansion amount in the overhanging direction is larger than the expansion amount in the vertical overhanging direction, and the length of the long axis is increased along with the increase of the overhanging length.
  8. 8. The intelligent sorting robot dynamic obstacle avoidance method based on multi-sensor fusion according to claim 6, wherein the hierarchical safety domain threshold in step S5 at least comprises a deceleration threshold and a parking threshold, the threshold is calculated based on time-to-collision TTC and/or minimum safety distance and at least one of the deceleration threshold and the parking threshold monotonically increases when the overhang length increases, and the threshold adjustment is implemented by a linear function or a piecewise linear function and is constrained by a preset safety range.
  9. 9. The intelligent sorting robot dynamic obstacle avoidance method based on multi-sensor fusion of claim 6, wherein the list of targets includes a target category field, and wherein a more conservative threshold adjustment factor and/or a greater expansion is employed for pedestrian targets than for drop packages in step S5 to make conditions for the pedestrian targets to trigger a deceleration or parking domain more stringent.

Description

Intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion Technical Field The invention belongs to the technical field of robot safety control and obstacle avoidance, and particularly relates to an intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion. Background When the mobile chassis sorting robot performs operations such as package pushing and sorting, diversion and the like in a sorting area, obstacle sensing and obstacle avoidance control are generally realized by relying on information of a visual sensor, a distance sensor and an odometer. In the prior art, the local environment occupation expression and the safety domain threshold are established based on the vehicle body fixed shape envelope. When the push-off/flow-guiding executing mechanism is in an extending, swinging or multi-pose state, the robot appearance envelope changes along with the push-off/flow-guiding executing mechanism, and the occupation expression under the fixed envelope is inconsistent with the threshold judgment and the real collision boundary, so that the following objective problems are caused: The method comprises the steps of (1) synchronously adjusting the triggering conditions of a safety domain along with the appearance change of an actuating mechanism, wherein hysteresis risks exist in deceleration or parking triggering, (2) occupying a grid to not embody a sweep boundary introduced by the posture of the actuating mechanism, and not expressing the real appearance constraint in an obstacle avoidance planning space, (3) when a fixed threshold is overlapped with a fixed envelope in a dynamic obstacle dense environment, causing error parking due to insufficient or excessive conservation of a collision margin, and not enough operation beat stability. Aiming at the problem that the robot profile envelope changes due to the gesture change of a push-separation/flow-guiding executing mechanism, so that the local occupation expression and the safety domain threshold judgment deviate from the real collision boundary, the invention establishes closed loops of executing mechanism gesture, profile envelope parameters, grid expansion, threshold tightening, hierarchical safety domain and linkage control, so that the safety domain judgment is simultaneously constrained by the gesture envelope occupation grid and gesture association threshold, and a linkage control instruction of a chassis and the executing mechanism is output. Disclosure of Invention The invention aims to provide an intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion, which is used for solving the technical problem that in the prior art, the gesture change of a pushing/guiding executing mechanism causes the change of the robot appearance envelope, so that the local occupation expression and the safety domain threshold judgment deviate from the real collision boundary. In order to achieve the above purpose, the present invention adopts the following technical scheme: An intelligent sorting robot dynamic obstacle avoidance system based on multi-sensor fusion, the sorting robot is used for executing pushing separation/diversion operation on packages, and comprises: The perception fusion unit is used for generating a target list of dynamic obstacles based on the environmental information of at least one visual sensor and at least one distance sensor and combining with the odometer information of the mobile chassis, and generating a local environment occupation grid; the gesture envelope parameter unit is used for acquiring gesture state information of the push-off/flow-guiding executing mechanism and determining outline envelope parameters according to the gesture state information; The gesture association safety domain unit is used for performing obstacle expansion processing on the local environment occupation grid according to the profile envelope parameters to obtain a gesture envelope occupation grid, adjusting a threshold value of the gesture association safety domain based on the profile envelope parameters, and judging an early warning domain, a deceleration domain and a parking domain according to the gesture envelope occupation grid and the adjusted threshold value; the linkage control unit is used for outputting a speed limiting instruction and/or a braking instruction of the mobile chassis according to the hierarchical security domain and outputting a locking instruction and/or a recycling instruction of the push-split/flow-guiding executing mechanism; The hierarchical security domain threshold tightens as the profile envelope parameter increases, and an amount of inflation of the obstacle inflation process is determined by the profile envelope parameter such that the hierarchical security domain decision is simultaneously constrained by the pose envelope occupancy grid and the pose association threshold. Further, the profile envelope parameters at least compri