Methods and systems for controlling a semiconductor fabrication process
First Claim
Patent Images
1. A method comprising:
- controlling operation of a semiconductor manufacturing system with a neural network to schedule processing of one or more workpieces, the semiconductor manufacturing system including at least two workpiece handling robots in at least two processing planes, the robots capable of handing workpieces between the at least two planes;
creating a data structure for one of the one or more workpieces, the data structure including an identity of the one or more workpieces and one or more fields for storing state information of the one or more workpieces; and
receiving data from the semiconductor manufacturing system as inputs to the neural network;
wherein the neural network is configured to effect modification of a processing task that changes a physical characteristic of a structure of the one or more workpieces in scheduled processes for each individual workpiece based on the state information related to the physical characteristic of the structure of respective ones of the one or more workpieces and the data received from the semiconductor manufacturing system; and
wherein the neural network outputs weights for one or more states of a finite state machine where the one or more states provide control signals for controlling operation of a semiconductor manufacturing system, and the neural network separates processing that reconfigures the state machine from operation of the finite state machine such that the finite state machine is enabled to operate independent of the neural network for allowing the neural network to evaluate outputs where processing time of the neural network extends beyond a time increment of the finite state machine.
8 Assignments
0 Petitions
Accused Products
Abstract
Software for controlling processes in a heterogeneous semiconductor manufacturing environment may include a wafer-centric database, a real-time scheduler using a neural network, and a graphical user interface displaying simulated operation of the system. These features may be employed alone or in combination to offer improved usability and computational efficiency for real time control and monitoring of a semiconductor manufacturing process. More generally, these techniques may be usefully employed in a variety of real time control systems, particularly systems requiring complex scheduling decisions or heterogeneous systems constructed of hardware from numerous independent vendors.
-
Citations
19 Claims
-
1. A method comprising:
-
controlling operation of a semiconductor manufacturing system with a neural network to schedule processing of one or more workpieces, the semiconductor manufacturing system including at least two workpiece handling robots in at least two processing planes, the robots capable of handing workpieces between the at least two planes; creating a data structure for one of the one or more workpieces, the data structure including an identity of the one or more workpieces and one or more fields for storing state information of the one or more workpieces; and receiving data from the semiconductor manufacturing system as inputs to the neural network; wherein the neural network is configured to effect modification of a processing task that changes a physical characteristic of a structure of the one or more workpieces in scheduled processes for each individual workpiece based on the state information related to the physical characteristic of the structure of respective ones of the one or more workpieces and the data received from the semiconductor manufacturing system; and wherein the neural network outputs weights for one or more states of a finite state machine where the one or more states provide control signals for controlling operation of a semiconductor manufacturing system, and the neural network separates processing that reconfigures the state machine from operation of the finite state machine such that the finite state machine is enabled to operate independent of the neural network for allowing the neural network to evaluate outputs where processing time of the neural network extends beyond a time increment of the finite state machine. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14)
-
-
15. A computer program product comprising computer executable code embodied in a non-transitory computer readable medium that, when executing on one or more computing devices, performs the steps of:
-
controlling operation of a semiconductor manufacturing system with a neural network to schedule processing of one or more workpieces, the semiconductor manufacturing system including at least two workpiece handling robots in at least two processing planes, the robots capable of handing workpieces between the at least two planes; creating a data structure for one of the one or more workpieces, the data structure including an identity of the one workpiece and one or more fields for storing state information of the one workpiece; receiving data from the semiconductor manufacturing system as inputs to the neural network; and modifying a processing task that changes a physical characteristic of a structure of the one or more workpieces in scheduled processes for each individual workpiece with the neural network based on the state information related to the physical characteristic of the structure of respective ones of the one or more workpieces and the data received from the semiconductor manufacturing system; and wherein the neural network outputs weights for one or more states of a finite state machine where the one or more states provide control signals for controlling operation of a semiconductor manufacturing system, and the neural network separates processing that reconfigures the state machine from operation of the finite state machine such that the finite state machine is enabled to operate independent of the neural network for allowing the neural network to evaluate outputs where processing time of the neural network extends beyond a time increment of the finite state machine.
-
-
16. A system comprising:
-
a hardware control unit including a neural network connected to and configured to control operation of a semiconductor manufacturing system to schedule processing of one or more workpieces, the semiconductor manufacturing system including at least two workpiece handling robots in at least two processing planes, the robots capable of handing workpieces between the at least two planes, wherein the neural network controller receives data from the semiconductor manufacturing system; and a hardware memory including a workpiece data structure, the data structure including an identity of one of the one or more workpieces and one or more fields for storing state information of the workpiece; wherein the neural network is configured to effect modification of a processing task that changes a physical characteristic of a structure of the one or more workpieces in scheduled processes for each individual workpiece based on the state information related to the physical characteristic of the structure of respective ones of the one or more workpieces and the data received from the semiconductor manufacturing system; and wherein the neural network outputs weights for one or more states of a finite state machine where the one or more states provide control signals for controlling operation of a semiconductor manufacturing system, and the neural network separates processing that reconfigures the state machine from operation of the finite state machine such that the finite state machine is enabled to operate independent of the neural network for allowing the neural network to evaluate outputs where processing time of the neural network extends beyond a time increment of the finite state machine. - View Dependent Claims (17, 18, 19)
-
Specification