IP Library › Granted Patent US 10,241,324
Granted Patent B2
US 10,241,324 · App. 15/839,217 · Granted Mar 26, 2019

Machine learning device for learning procedure for aligning optical part of light source unit, and light-source unit manufacturing apparatus

Inventor: Michinori Maeda (Yamanashi, JP)
Assignee: FANUC CORPORATION
G02B26/0875G02B27/0012G02B27/62G06F15/18
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 10,241,324
App. No.
15/839,217
Granted
Mar 26, 2019
Kind
B2
Abstract

A machine learning device acquires decision data including an adjustment time of a position and an orientation of an optical part and a state of light measured by a light measurement device. The machine learning device includes a learning unit that learns a procedure for adjusting the position and the orientation of the optical part. The learning unit includes a reward calculation unit that calculates a reward based on the alignment adjustment time and the state of light, and a value function updating unit that updates a value function based on the reward. The learning unit includes a decision unit that sets a movement method of the optical part based on the value function.

Claims (26)

1. A machine learning device that learns a procedure for aligning optical parts of a light source unit, the machine learning device comprising:

a state observation unit that acquires state data including a position and an orientation of the optical part and quality information on a component of the light source unit;

a decision data acquisition unit that acquires decision data including an adjustment time of the position and the orientation of the optical part and a state of light measured by a light measurement device; and

a learning unit that learns a procedure for adjusting the position and the orientation of the optical part based on the state data acquired by the state observation unit and the decision data acquired by the decision data acquisition unit,

the learning unit including:

a reward calculation unit that calculates a reward based on the adjustment time and the state of light that are acquired by the decision data acquisition unit;

a value function updating unit that updates a value function based on the reward set by the reward calculation unit; and

a decision unit that sets a movement method of the optical part based on the value function.

2. The machine learning device according to claim 1 , wherein the state of light includes at least one of light intensity, a mode pattern, and a center position of light.

3. The machine learning device according to claim 1 , wherein

the components are lenses, and

the quality information on the components includes at least one of material component, lens shape, index of refraction, focal length, characteristics of an optical thin film deposited on a surface, material lot, manufacturing apparatus, operator, manufacturing process, and date of manufacture.

4. The machine learning device according to claim 1 , further comprising a communication device that communicates with the other machine learning device; wherein

the machine learning device receives state data and decision data from the other machine learning device, and shares the value function with the other machine learning device.

5. A light-source unit manufacturing apparatus comprising:

a light measurement device that measures a state of light outputted from a light-emitting element;

a movement device that moves an optical part while holding the optical part; and

a controller that controls the movement device,

the controller including:

a state observation unit that acquires state data including a position and an orientation of the optical part and quality information on a component;

a decision data acquisition unit that acquires decision data including an adjustment time of the position and the orientation of the optical part and the state of light measured by the light measurement device; and

a learning unit that learns a procedure for adjusting the position and the orientation of the optical part based on the state data acquired by the state observation unit and the decision data acquired by the decision data acquisition unit,

the learning unit including:

a reward calculation unit that calculates a reward based on the adjustment time and the state of light that are acquired by the decision data acquisition unit;

a value function updating unit that updates a value function based on the reward set by the reward calculation unit; and

a decision unit that sets a movement method of the optical part based on the value function.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Dec 12, 2017
From: MAEDA, MICHINORI
To: FANUC CORPORATION
Reel/Frame 044372/0191 →
Priority Claims (1)
JP 2016-245834 · Dec 19, 2016 · national
Continuity (1)
Related Publication 20180172983A1 · Jun 21, 2018