SAFETY REAL-TIME OPTIMAKE

机械臂在线轨迹规划 Online Trajectory Planning for Manipulators

面向码垛、上下料等场景,在动态工位中实时生成安全且最优的运动轨迹。 For palletizing and machine tending: real-time generation of safe, optimal motions in dynamic workcells.

Motion Planning

码垛与上下料现场,需要会“绕开”的轨迹 Palletizing and Machine Tending Need Motions That Know How to Avoid

任务场景The Task

码垛、拆垛、上下料的工位从来不是空的:垛型在长高、周转箱在移动、人员偶尔进入。机械臂需要在这样的环境中连续生成安全、可执行且高效的轨迹。

  • 避开动态障碍:垛型与周转箱实时变化,示教轨迹无法覆盖所有工况。
  • 守住安全边界:设备、围栏与人员区域必须成为不可逾越的限制。
  • 兼顾节拍与能耗:既要缩短单次搬运时间,也要控制电机发热与能耗。

Palletizing, depalletizing, and machine-tending cells are never empty: stacks grow, totes move, and people occasionally step in. The manipulator must continuously generate motions that are safe, executable, and efficient.

  • Avoid moving obstacles: stacks and totes change constantly; taught paths cannot cover every case.
  • Respect safety bounds: equipment, fences, and human zones must be hard limits.
  • Balance cycle time and energy: cut handling time while keeping motor heating and energy use in check.
问题抽象Problem Formulation
min time or energy
s.t. kinematics
collision avoidance
initial and terminal pose

现场需求可以直接写成一个带约束的优化问题:以节拍或能耗为目标,把运动学、避障与起终点位姿作为约束,由求解器统一处理。

The requirements above map directly onto a constrained optimization problem: minimize cycle time or energy, subject to kinematics, collision avoidance, and boundary poses — all handled by the solver at once.

Palletizing motion planning
OCC
动态占据栅格Occupancy grid
Boundary
虚拟围栏Virtual fence
ms
在线重规划Online replanning
Obstacle avoidance and virtual fence
01 Safety

避障与围栏作为硬约束,天然写进优化问题 Obstacle and Fence Limits Enter the Problem as Hard Constraints

OPTIMake直接支持避障与安全约束的建模:动态占据栅格(OCC)、几何包络、虚拟围栏(boundary)等避障条件可直接作为约束加入规划问题,由求解器保证满足,而不是事后检查再重试。

垛型变化、料箱移动、人员进入安全区,都会即时反映为新的约束,轨迹自动绕行而非急停。

OPTIMake supports modeling of obstacle-avoidance and safety constraints directly: dynamic occupancy grids (OCC), geometric envelopes, virtual fences (boundaries), and similar conditions can be added straight into the planning problem as constraints and are enforced by the solver, rather than checked afterwards and retried.

Changing stacks, moving totes, and people entering a safety zone immediately become new constraints, so the robot detours instead of emergency-stopping.

02 Real-Time

毫秒级在线重规划,环境一变立刻响应 Millisecond Replanning Responds the Moment the Scene Changes

规划在毫秒量级完成,可以跟随感知刷新率持续重算,无需依赖离线示教或预存轨迹库。

新垛位、新工件尺寸、临时插入的障碍都无需重新示教,产线切换更快。

Planning completes in milliseconds and can be recomputed at the perception refresh rate, without offline teaching or stored trajectory libraries.

New pallet positions, part sizes, or temporary obstacles require no re-teaching, making changeover faster.

03 Optimality

以时间或能耗为目标,直接求最优而非可行 Optimize for Time or Energy Instead of Settling for Feasible

规划目标可直接设为最短时间或最小能耗,并在同一问题中兼顾平滑性,输出的是最优轨迹而不只是一条可行路径。

节拍与能耗之间的权衡只需调整目标权重,方便针对不同产线诉求快速标定。

The objective can be set directly to minimum time or minimum energy while accounting for smoothness in the same problem, producing an optimal motion rather than merely a feasible one.

Trading cycle time against energy is a matter of adjusting objective weights, making it easy to tune for different production goals.

想在您的机器人上验证效果? Want to validate this on your robot?

我们提供试用版本、演示案例与技术支持 We provide trial versions, demo cases, and technical support.

联系我们Contact us