Toolpath generation by reinforcement learning for computer aided manufacturing
Methods, systems, and apparatus, including medium-encoded computer program products, for computer aided design and manufacture of physical structures using toolpaths generated by reinforcement learning for use with subtractive manufacturing systems and techniques, include: obtaining, in a computer aided design or manufacturing program, a three dimensional model of a manufacturable object; generating toolpaths that are usable by a computer-controlled manufacturing system to manufacture at least a portion of the manufacturable object by providing at least a portion of the three dimensional model to a machine learning algorithm that employs reinforcement learning, wherein the machine learning algorithm includes one or more scoring functions that include rewards that correlate with desired toolpath characteristics comprising toolpath smoothness, toolpath length, and avoiding collision with the three dimensional model; and providing the toolpaths to the computer-controlled manufacturing system to manufacture at least the portion of the manufacturable object.
1 . A system comprising:
a data processing apparatus including at least one hardware processor; and
a non-transitory computer-readable medium encoding instructions configured to cause the data processing apparatus to perform operations comprising:
obtaining a three dimensional model of a manufacturable object;
generating toolpaths that are usable by a computer-controlled manufacturing system to manufacture at least a portion of the manufacturable object by providing at least a portion of the three dimensional model to two or more machine learning algorithms, wherein the generating comprises
generating data comprising image feature data or at least one starting position for the toolpaths by processing the portion of the three dimensional model with a first of the two or more machine learning algorithms, the image feature data or the at least one starting position being input for creation of the toolpaths, and
generating the toolpaths by processing the data comprising the image feature data or the at least one starting position for the toolpaths with a second of the two or more machine learning algorithms, wherein at least the second of the two or more machine learning algorithms employs reinforcement learning including one or more scoring functions that include rewards that correlate with one or more desired toolpath characteristics, and the second of the two or more machine learning algorithms is different from the first of the two or more machine learning algorithms; and
providing the toolpaths to the computer-controlled manufacturing system to manufacture at least the portion of the manufacturable object.
2 . The system of claim 1 , wherein the generating comprises:
processing a global view of the portion of the three dimensional model with the first of the two or more machine learning algorithms to generate the data comprising the at least one starting position for the toolpaths; and
processing a local view of the portion of the three dimensional model at least around each of the at least one starting position with the second of the two or more machine learning algorithms to generate the toolpaths near each of the at least one starting position.
3 . The system of claim 2 , wherein the global view comprises a low resolution representation of at least the portion of the three dimensional model, and the local view comprises a high resolution representation of the portion of the three dimensional model at least around each of the at least one starting position.
4 . The system of claim 1 , wherein the first of the two or more machine learning algorithms comprises a convolutional neural network, and the data comprises the image feature data extracted by the convolutional neural network from one or more images representing one or more two dimensional cross sections of at least the portion of the three dimensional model.
5 . The system of claim 1 , wherein the operations further comprise:
defining the one or more scoring functions that include the rewards that correlate with the one or more desired toolpath characteristics;
receiving a plurality of training examples, each training example comprising example tools and example environments; and
training the second of the two or more machine learning algorithms that employs reinforcement learning using the training examples to generate toolpaths that maximize one or more values generated by the one or more scoring functions.
6 . The system of claim 1 , wherein the two or more machine learning algorithms generate the toolpaths comprising an angle for a cutter used by the computer-controlled manufacturing system.
7 . The system of claim 1 , wherein the second of the two or more machine learning algorithms generates a series of locations for a tool used by the computer-controlled manufacturing system, the toolpaths comprise one or more splines that connect the series of locations for the tool, and providing the toolpaths to the computer-controlled manufacturing system to manufacture at least the portion of the manufacturable object comprises providing the one or more splines to control the tool to travel through the series of locations in a smooth route following the one or more splines.
8 . The system of claim 1 , wherein the second of the two or more machine learning algorithms operates in a three dimensional environment and uses a three dimensional representation of a tool used by the computer-controlled manufacturing system.
9 . The system of claim 1 , wherein the one or more desired toolpath characteristics comprise a desired side of a tool in the toolpaths.
10 . The system of claim 1 , wherein the one or more desired toolpath characteristics comprise rotation characteristics of a tool.
11 . The system of claim 1 , wherein the one or more desired toolpath characteristics comprise tool engagement for a selected cutting tool.
12 . The system of claim 1 , wherein the second of the two or more machine learning algorithms employs a percentage of cutter used.
13 . The system of claim 1 , wherein the system comprises the computer-controlled manufacturing system.
14 . The system of claim 13 , wherein the computer-controlled manufacturing system is a multi-axis, multi-tool milling machine.
15 . The system of claim 1 , wherein the two or more machine learning algorithms are trained to generate toolpaths for 2.5-axis machining.
16 . The system of claim 15 , wherein the three dimensional model is obtained using a generative design software that generates the three dimensional model including a plurality of discrete layers.
17 . A non-transitory computer-readable medium encoding instructions operable to cause a data processing apparatus to perform operations comprising:
obtaining a three dimensional model of a manufacturable object;
generating toolpaths that are usable by a computer-controlled manufacturing system to manufacture at least a portion of the manufacturable object by providing at least a portion of the three dimensional model to two or more machine learning algorithms, wherein the generating comprises
generating data comprising image feature data or at least one starting position for the toolpaths by processing the portion of the three dimensional model with a first of the two or more machine learning algorithms, the image feature data or the at least one starting position being input for creation of the toolpaths, and
generating the toolpaths by processing the data comprising the image feature data or the at least one starting position for the toolpaths with a second of the two or more machine learning algorithms, wherein at least the second of the two or more machine learning algorithms employs reinforcement learning including one or more scoring functions that include rewards that correlate with one or more desired toolpath characteristics, and the second of the two or more machine learning algorithms is different from the first of the two or more machine learning algorithms; and
providing the toolpaths to the computer-controlled manufacturing system to manufacture at least the portion of the manufacturable object.
18 . The computer-readable medium of claim 17 , wherein the generating comprises:
processing a global view of the portion of the three dimensional model with the first of the two or more machine learning algorithms to generate the data comprising the at least one starting position for the toolpaths; and
processing a local view of the portion of the three dimensional model at least around each of the at least one starting position with the second of the two or more machine learning algorithms to generate the toolpaths near each of the at least one starting position.
19 . The computer-readable medium of claim 17 , wherein the first of the two or more machine learning algorithms comprises a convolutional neural network, and the data comprises the image feature data extracted by the convolutional neural network from one or more images representing one or more two dimensional cross sections of at least the portion of the three dimensional model.
20 . The computer-readable medium of claim 18 , wherein the global view comprises a low resolution representation of at least the portion of the three dimensional model, and the local view comprises a high resolution representation of the portion of the three dimensional model at least around each of the at least one starting position.
21 . The computer-readable medium of claim 17 , wherein the operations further comprise:
defining the one or more scoring functions that include the rewards that correlate with the one or more desired toolpath characteristics;
receiving a plurality of training examples, each training example comprising example tools and example environments; and
training the second of the two or more machine learning algorithms that employs reinforcement learning using the training examples to generate toolpaths that maximize one or more values generated by the one or more scoring functions.
22 . The computer-readable medium of claim 17 , wherein the two or more machine learning algorithms generate the toolpaths comprising an angle for a cutter used by the computer-controlled manufacturing system.
23 . The computer-readable medium of claim 17 , wherein the second of the two or more machine learning algorithms generates a series of locations for a tool used by the computer-controlled manufacturing system, the toolpaths comprise one or more splines that connect the series of locations for the tool, and providing the toolpaths to the computer-controlled manufacturing system to manufacture at least the portion of the manufacturable object comprises providing the one or more splines to control the tool to travel through the series of locations in a smooth route following the one or more splines.
24 . The computer-readable medium of claim 17 , wherein the second of the two or more machine learning algorithms operates in a three dimensional environment and uses a three dimensional representation of a tool used by the computer-controlled manufacturing system.
25 . The computer-readable medium of claim 17 , wherein the one or more desired toolpath characteristics comprise a desired side of a tool in the toolpaths.
26 . The computer-readable medium of claim 17 , wherein the one or more desired toolpath characteristics comprise rotation characteristics of a tool.
27 . The computer-readable medium of claim 17 , wherein the one or more desired toolpath characteristics comprise tool engagement for a selected cutting tool.
28 . The computer-readable medium of claim 17 , wherein the second of the two or more machine learning algorithms employs a percentage of cutter used.
29 . The computer-readable medium of claim 17 , wherein the computer-controlled manufacturing system is a multi-axis, multi-tool milling machine.
30 . The computer-readable medium of claim 17 , wherein the two or more machine learning algorithms are trained to generate toolpaths for 2.5-axis machining.
31 . The computer-readable medium of claim 30 , wherein the three dimensional model is obtained using a generative design software that generates the three dimensional model including a plurality of discrete layers.
32 . A method comprising:
obtaining a three dimensional model of a manufacturable object;
generating toolpaths that are usable by a computer-controlled manufacturing system to manufacture at least a portion of the manufacturable object by providing at least a portion of the three dimensional model to two or more machine learning algorithms, wherein the generating comprises
generating data comprising image feature data or at least one starting position for the toolpaths by processing the portion of the three dimensional model with a first of the two or more machine learning algorithms, the image feature data or the at least one starting position being input for creation of the toolpaths, and
generating the toolpaths by processing the data comprising the image feature data or the at least one starting position for the toolpaths with a second of the two or more machine learning algorithms, wherein at least the second of the two or more machine learning algorithms employs reinforcement learning including one or more scoring functions that include rewards that correlate with one or more desired toolpath characteristics, and the second of the two or more machine learning algorithms is different from the first of the two or more machine learning algorithms; and
providing the toolpaths to the computer-controlled manufacturing system to manufacture at least the portion of the manufacturable object.
33 . The method of claim 32 , wherein the generating comprises:
processing a global view of the portion of the three dimensional model with the first of the two or more machine learning algorithms to generate the data comprising the at least one starting position for the toolpaths; and
processing a local view of the portion of the three dimensional model at least around each of the at least one starting position with the second of the two or more machine learning algorithms to generate the toolpaths near each of the at least one starting position.
34 . The method of claim 33 , wherein the global view comprises a low resolution representation of at least the portion of the three dimensional model, and the local view comprises a high resolution representation of the portion of the three dimensional model at least around each of the at least one starting position.
35 . The method of claim 32 , wherein the first of the two or more machine learning algorithms comprises a convolutional neural network, and the data comprises the image feature data extracted by the convolutional neural network from one or more images representing one or more two dimensional cross sections of at least the portion of the three dimensional model.
36 . The method of claim 32 , further comprising:
defining the one or more scoring functions that include the rewards that correlate with the one or more desired toolpath characteristics;
receiving a plurality of training examples, each training example comprising example tools and example environments; and
training the second of the two or more machine learning algorithms that employs reinforcement learning using the training examples to generate toolpaths that maximize one or more values generated by the one or more scoring functions.
37 . The method of claim 32 , wherein the two or more machine learning algorithms generate the toolpaths comprising an angle for a cutter used by the computer-controlled manufacturing system.
38 . The method of claim 32 , wherein the second of the two or more machine learning algorithms generates a series of locations for a tool used by the computer-controlled manufacturing system, the toolpaths comprise one or more splines that connect the series of locations for the tool, and providing the toolpaths to the computer-controlled manufacturing system to manufacture at least the portion of the manufacturable object comprises providing the one or more splines to control the tool to travel through the series of locations in a smooth route following the one or more splines.
39 . The method of claim 32 , wherein the second of the two or more machine learning algorithms operates in a three dimensional environment and uses a three dimensional representation of a tool used by the computer-controlled manufacturing system.
40 . The method of claim 32 , wherein the one or more desired toolpath characteristics comprise a desired side of a tool in the toolpaths.
41 . The method of claim 32 , wherein the one or more desired toolpath characteristics comprise rotation characteristics of a tool.
42 . The method of claim 32 , wherein the one or more desired toolpath characteristics comprise tool engagement for a selected cutting tool.
43 . The method of claim 32 , wherein the second of the two or more machine learning algorithms employs a percentage of cutter used.
44 . The method of claim 32 , wherein the computer-controlled manufacturing system is a multi-axis, multi-tool milling machine.
45 . The method of claim 32 , wherein the two or more machine learning algorithms are trained to generate toolpaths for 2.5-axis machining.
46 . The method of claim 45 , wherein the three dimensional model is obtained using a generative design software that generates the three dimensional model including a plurality of discrete layers.