Method: Vision based Tool Retraction Until Safe Tool-path (V- TRUST)

General context and approach

In conventional multi-axis machining production the Computer- Aided  Design  and  Manufacturing  (CAD/CAM)  process   program

Fig. 2. 3D-MLI sensor installed on the milling machine. (a) 3D MLI sensor and (b) manufacturing scene.

a desired tool-path for manufacturing a part, which is transferred to the real machine using a postprocessor. In the case of collisions or any change at the shop floor, the whole programing process needs to be restarted because of the  feeble  feedback  system and unidirectional post-processor nature. In order to give more autonomy to the machine for decision regarding safe path planning, it is important to treat the real information from the scene and decide online about collision detection and avoidance at machining  level.

A ToF sensor based 3D information collection system and decision method is devised on multi-axis machine for collision detection and avoidance as shown in Fig. 3. The Vision based Tool Retraction Until Safe Tool-path (V-TRUST) method presented in this paper is based on 3D image information extraction and processing by using a 3D-MLI sensor on a real machining scenario. The sensor exacts 3D information about the obstacle and using V- TRSUT methodology will draw an efficient decision to detect and avoid collisions by shifting tool to a secure position and ultimately to  the final goal production point   location.

The selection of a secure position is critical and that is the one of the important thing which this paper proposes different from the previous approaches. The secure point should facilitate tool to reach the final goal point by taking into account the architecture of machine for easy accessibility with minimum time and having no further collision. The method presented in this paper will take information about the initial programed trajectory from the CAD/ CAM program. This interpreted trajectory is required to know the initial planning of the production tool. While exploiting the information from the real scene and the trajectory interpretation data, the method proposed will decide secure trajectories for the production tool. The best trajectory is selected using several criteria such as minimum length and time. The final safe trajectory planned is then send to the machine where collision is avoided. The final safe trajectory is then transferred back to the machine either off-line or online. Currently the construction of multi-axis machine based on conventional G-code does not permit to do this online and therefore the method proposed is restricted to the off-line safe trajectory solution in 3D with real machine     information.

In future the same method will be applied to real-time (online) and dynamic safe trajectory generation, when the new generation of high level STEP-NC based controllers [27] will be available.

Application scenario

A production tool tip point is considered for tool-path generation as  an  initial  application,  whereas  tool-path  in  focus is rapid transversal, non-functional tool-path between two functional points. Fig. 4 shows a real machine scenario in focus, where a production tool has to move from reference position (point

1) to two different workpieces (points 2 and 4) for drilling and an unknown, un-programed obstacle may cause collision (point 3). Many reasons are highlighted in our previous work [9] for the presence of collisions in multi-axis machines e.g. difference between real machining and its virtual simulations, operators’ mistake of not correctly entering the dimensions of various objects in the scene, broken piece  of machine, misplaced fixture,     etc.

V-TRUST  algorithm

The flowchart of V-TRUST algorithm for safe tool-path is shown in Fig. 5. Actual 3D information from the manufacturing scene is taken in the presence and absence of obstacle using 3D-MLI sensor. While knowing different manufacturing points as shown in Fig. 4, the approach presented will calculate tool-path from initial point to final goal point using interpreted trajectory according to the machine strategy, and thereby verifying each next tool position for anticipated  collision. 铣削和钻削数控机床英文文献和中文翻译(3):http://www.chuibin.com/fanyi/lunwen_206129.html