US-12619243-B1 - Machine learning based reconfigurable mobile agents using swarm system manufacturing

Abstract

Embodiments of the present invention include a collaborative manufacturing system utilizing a plurality of mobile agents. The mobile agents operate dual robotic arms to improve single and multi-material builds' efficiency. In some embodiments, the dual robotic arms work together in the same area to create multi-functional components. In addition, the mobile agents can change tool heads on the arms to allow for hybrid manufacturing such as pick and place, additive, and subtractive manufacturing. One or more mobile agents interact with other mobile agents, thereby increasing the end product's efficiency and quality. Mobile agents utilize swarm manufacturing techniques to improve the manufactured product's time efficiency further and use machine learning to adjust and re-assign mobile agents constantly.

Inventors

• Tarik J. Dickens
• Jolie Breaux Frketic
• Sean Psulkowski

Assignees

• FLORIDA A&M UNIVERSITY

Dates

Publication Date

20260505

Application Date

20231030

Claims (5)

1. 1 . A system for reducing a manufacturing time of an end product within a predefined manufacturing area, the system comprising: an independent remote central controller configured to control the manufacturing of the end product within the predefined manufacturing area, the independent remote central controller in communication with a first mobile agent and a second mobile agent; the first mobile agent disposed within the predefined manufacturing area and configured to receive a first set of instructions transmitted by the independent remote central controller, the first mobile agent including a build platform coupled to a frame of the first mobile agent, the build platform having a surface configured to support an end product thereon, wherein the build platform comprises a print bed with an adhesive layer on the surface, the adhesive layer being configured to prevent the end product from being displaced during manufacture or movement of the first mobile agent, the second mobile agent, or both; the second mobile agent disposed within the predefined manufacturing area and configured to receive a second set of instructions transmitted by the independent remote central controller, the second mobile agent including an attachment having an end effector disposed at a terminal end of the attachment, the end effector configured to interact with the build platform of the first mobile agent based on the second set of instructions received by the second mobile agent to assist in a construction of the end product on the build platform of the first mobile agent; and wherein the independent remote central controller is configured to collect data from the end product via at least one sensing component on the first mobile agent, the second mobile agent or both, and whereby the independent remote central controller processes the data to evaluate the performance of the end product for a subsequent end product, thereby reducing the overall manufacturing time.
2. 2 . The system of claim 1 , wherein the first mobile agent includes: an ultrasound detector in mechanical communication with the frame, the ultrasound detector being configured to calculate a void content of the first end product.
3. 3 . The system of claim 1 , wherein the first mobile agent further comprises a camera in mechanical communication with the frame of the first mobile agent and is configured to capture video data within the manufacturing area, wherein the video data is wirelessly transmitted to the independent remote central controller.
4. 4 . The system of claim 1 , wherein the attachment of the second mobile agent is a robotic arm having a motor configured to control the movement of the robotic arm within the manufacturing area.
5. 5 . The system of claim 1 , wherein each of the first and the second mobile agents includes omni-directional wheelings coupled to the frame, the omni-directional wheelings configured to permit each of the first and the second mobile agents to independently linearly translate within the manufacturing area.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS This nonprovisional application is a divisional of and claims the benefit of U.S. Nonprovisional patent application Ser. No. 17/247,300 entitled “MACHINE LEARNING BASED RECONFIGURABLE MOBILE AGENTS USING SWARM SYSTEM MANUFACTURING” filed Dec. 7, 2020 by the same inventors, which claims priority to U.S. Provisional Patent Application No. 62/944,871 entitled “MACHINE LEARNING BASED RECONFIGURABLE MOBILE AGENTS USING SWARM SYSTEM MANUFACTURING” filed Dec. 6, 2019 by the same inventors, all of which are incorporated herein by reference, in their entireties, for all purposes. BACKGROUND OF THE INVENTION 1. Field of the Invention This invention relates, generally, to additive and subtractive manufacturing systems. More specifically, it relates to reconfigurable mobile agents utilizing swarm manufacturing techniques to reduce overall manufacturing time and increase the quality of an end product. 2. Brief Description of the Prior Art Significant drawbacks exist in existing gantry style equipment. First, traditional gantry style manufacturing equipment (GSME) is stationary, lacking the mobility to move around the manufacturing area. This limitation is compounded by the large blueprint and heavy servo motors of traditional GSME. Second, GSME includes large support columns and gantry equipment that restricts the movement of Selective Compliance Assembly Robot Arm (hereinafter “SCARA”) movement within a build area. Third, GSME cannot quickly swap out end effector toolings, which increases the overall build time of a fabricated part (alternatively referred to as “end product”) since end effectors must be manually interchanged. These and other drawbacks readily apparent to one of ordinary skill in the art result in significant inefficiencies during the manufacturing process. With the introduction of precision additive manufacturing techniques, the complexity of manufactured end products rivals the possibilities of conventional manufacturing alternatives. However, the time investment required to bring an end product to completion is significant. Furthermore, this time commitment severely limits the practicality of an otherwise advantageous shift in manufacturing. Additionally, with the recent paradigm shift to rapidly manufacturing unconventional end products for Industry 4.0 (i.e., the ongoing automation of traditional manufacturing and industrial practices, also referred to as the Fourth Industrial Revolution) and the factory of the future, the industrial Additive Manufacturing (AM) complex continues an upward trajectory that will be worth an estimated $35.6 billion by 2024 [1]. This increase in popularity of AM is primarily due to the pervasive access to advanced manufacturing for personalization, customization, and product design turnover of future markets-commonly referred to as “agile manufacturing” [2]. Largely the most popular mode of AM, fused filament fabrication and modeling (FFF/M), demonstrates a high potential for fabricating multi-functional 3D-printed parts on-par with conventional processing methods. Moreover, FFF/M can grow amongst a variety of applications, heavily driven by a comprehensive material library of available materials [3] and ease in development [4]. In particular, the additive process is enabled through filament material fed via a motor through a heated nozzle, which then constructs sequential layers ending in a final end product geometry. Overall, the benefits inherent to the FFF/M process consist of (i) reduced material waste, (ii) flexibility in application deployment, and (iii) growing popularity among industry and makerspaces. While many AM projects are completed via a singular machine, multiple devices can work in conjunction to complete a finalized end product. For example, collaborative or cooperative robotics is a field in which multiple robotic/humanoid systems work together to achieve a goal through planning and deliberation between the robots through local communication and coordination [5-8]. Recent works show a rise in decentralized multi-agent systems for both extrusion and hybrid additive manufacturing. Increasing the number of mobile agents during manufacturing facilitates larger and more complex end products in less time than convention means [9-13]. Similarly, contemporary assembly line manufacturing utilizes Multi-Degree of Freedom (MDoF) manipulators, which provides greater agility in smaller operational space than their gantry predecessors, extending to use in modular assembly lines [14]. The benefits in efficiency and precision from using MDoF manipulators largely outweigh the limitations in complex path planning, as industrial mobile agents can operate for thousands of continuous hours without positional error. Attempts to migrate manufacturing processes (both additive and subtractive) from GSME to MDoF manipulators have proved nearly equivalent at the production scale. Thus, future manufacturing relies upon a modular assembly