Projects.

Check out some of my professional engineering highlights below!

Background

  • Micropsi Industries

    September 2023 - Present, San Francisco, CA

    Application Engineer, leading the US engineering team from proof of concept through product deployment. Designed solutions to automate hardware manufacturing and assembly processes, working with R&D teams from top tech and automotive companies.

  • Garmin

    July 2022 - August 2023, Olathe, KS

    Automation Process Engineer, led mechanical design efforts for an automated electronics hardware kitting work cell, and wrote robot code for automated electronics manufacturing processes. Before my full time position, I interned at Garmin for two summers in 2020 & 2021.

  • Iowa State University

    Fall 2017 - Spring 2022, Ames, IA

    BS Major in Mechanical Engineering, Minor in Computer Science.

Photo courtesy of micropsi industries

Cable Harness Pick & Plug Automation

  • Cable harnesses are collections of wires, connectors, and terminals which are routed through a car’s interior to carry power to the car’s electrical components. In this project, I was tasked with using MIRAI (AI vision software) and a FANUC robot to automate the picking & plugging of these cable harness connectors as they dangled from the bare-internals of a car door! In automotive manufacturing, automating this task with traditional methods is nearly impossible.

    • Gripper finger design

    • Camera mount design

    • Prototyping & fabricating setup

    • Training AI software policies for locating the cables

    • Robot programming

    • Demo design

    • Machine assembly

    • Assist marketing team in filming the demo

  • Gripping the different cable form factors with the same toolset and dealing with cable tension, were the two main challenges in this project. The differences in cable sizes & locating features initially led to complicated, puzzle-like gripping features. Although this strategy worked in a static execution, it did not work when the robot was holding & moving the parts—as tension in the wires would often force the cables out of the robots grip. So I switched gripping strategies, and I replaced the locating features with flat, rubber pads. Additionally, I shortened the gripper finger length to lessen the amount of deflection when gripping the parts. Together with these changes, there was enough friction created in the grip to prevent the cables from slip or moving under cable tension; and although the grip position was inconsistent without the locating features, I knew the AI vision software (MIRAI) would be able to compensate the robot positioning despite these inconsistencies. In training the AI policies, challenges included: lighting variation & direct sunlight, rotational variation in cable positioning, and visual noise created by the mess of dangling cables.

  • I overcame the gripper design challenges and was able to successfully train MIRAI policies to grab the different dangling cables out of free-space and plug them into the correct ports. This demo led to increased marketing engagement as well as new sales leads with automotives for this particular application.

What is MIRAI?

MIRAI is AI vision software for robot motion control, and it is the flagship product of the Berlin-based startup Micropsi Industries. Through my time with Micropsi, I’ve had the opportunity to become an expert user with this amazing tool, and to engineer novel automation solutions with some of the worlds top tech, manufacturing, and automotive companies. Learn more about MIRAI and micropsi industries below!

Micropsi's website

Although I can’t share everything, here’s a list of brief descriptions outlining past projects from 2019 through 2024.

Project History.

2024

  • Micropsi Industries

    Using AI vision + robotic arm to automate the insertion of FPCs in electronics assembly. I deployed MIRAI in Suzhou, China for this project.

  • Micropsi Industries

    Automation of an automotive general assembly process—marrying two large components of an electric vehicle, by guiding two robot arms with AI vision to perform bolt torquing task.

  • Micropsi Industries

    Proof of concept project, demonstrating the ability of AI vision to guide a very large robot in the depalletization of tires. I designed camera mounts for the robot EOAT, reviewed the customers gripper designs, and trained the AI policies to perform the tire-picking task.

  • Micropsi Industries

    Created a marketing demo, showing off MIRAIs ability to pick & plug cables dangling from the inside of a car door—a task that would be extremely difficult to automate with any other existing solution.

  • Micropsi Industries

    Designed robot EOAT & camera mounts so that MIRAI could automate cable plugging at an end-of-line testing station.

  • Micropsi Industries

    Automated the insertion of a “tinsel" wire into a tight fitting hole—with less than .20 mm of clearance. Achieved >99.2% success rate using MIRAI. Designed camera mounts, trained the AI policies to perform the task, and trained the customer on best-use practices for MIRAI.

  • Micropsi Industries

    Proof of concept using MIRAI to align the nozzle of a wax-dispenser to a tight fitting hole. Reviewed & assisted customer with EOAT design, trained AI policies to perform the task, presented results to customer team, and scouted the factory for additional MIRAI applications.

  • Micropsi Industries

    Created a tradeshow demo, showing off MIRAIs ability to guide robots in real time for high-variance, precise tasks—in this case, picking a flexible cable out of free-space and plugging it into a tight fitting port. I designed, built, programmed the entire demo, and trained all of the MIRAI policies to perform the different tasks. The demo was super reliable, and scaled to multiple machines for simultaneous showcasing throughout the year.

  • Micropsi Industries

    Demonstrated MIRAIs ability to pick-and-place parts which are presented in conditions of high variance—in this case, shiny metal disc-brakes presented in a malleable cardboard box. The demo was extremely reliable, and the MIRAI policies did not fail in 100+ hours of operation across the year’s 5 tradeshows.

2023

  • Automated the pick and insertion of FPCs. This was a lab proof-of-concept with an electronics hardware manufacturing group. I designed the camera mount setup for MIRAI, installed MIRAIs hardware in the lab, trained the AI policies for the various tasks, and trained the customer how to use MIRAI. The POC was a success, with the policies performing at >99% success rates. The project progressed into full deployment (see 2024 projects).

  • Micropsi Industries

    Automated bolt torquing tasks for hardware assembly. Deployed multiple MIRAI systems in customer facilities, and supported the customer by reviewing MIRAI policy data in order to meet production demands. Trained the customer on how to make their own MIRAI policies for other bolt torquing applications. The customer scaled MIRAI from 1 station to 4, and then from 1 process to 3 different processes.

  • Started a documentation rewrite initiative within my first 3 months at micropsi. I rewrote sections of the MIRAI Training user manual, and created diagrams & illustrations to teach customers how to use the product. My contributions are still featured in the user manuals.

  • Micropsi Industries

    Deployed MIRAI for the automation of a tig-welding station. Installed MIRAI hardware in the customer’s factory, trained MIRAI policies, and trained the customer engineer to use MIRAI.

  • Micropsi Industries

    In this proof of concept, I trained MIRAI policies to pick and maneuver dangling seatbelts—for the purpose of automotive hardware assembly at the supplier level.

  • Garmin

    Assembled all components I was responsible for, and volume tested them to ensure process reliability.

2022

  • Garmin

    Project Jill: a kitting automation work-cell for assembling and bagging a large variety of small parts (screws, washers, hardware casings, c-clips).

    • Machine design: I was project design lead for robot tooling, tool racking, assembly nest, fixturing, and pneumatics solutions

    • Designed a multi-pneumatic-actuator, c-clip inserter mechanism for accurate and repeatable part assembly

    • Designed multiple gripper finger sets for repeatable picking/placing tasks on a variety of parts

    • Designed table-mounted tool rack and tool interface plates for simple and precise automated robot tool changing

    • Created large CAD assemblies in SolidWorks; Prototyped with rapid fabrication resources (3D printing, machining, etc.)

Undergrad

  • My senior design team is helping Iowa State's Cardinal Space Mining Club get ahead of its competition! The club's current space mining robot specializes in regolith (moon dust) collection and trenching activities for NASA's annual Lunabotics competition. In anticipation of upcoming rule changes for next year's competition, my team is designing a robotic system that creates and lays regolith bricks in simulated lunar conditions.

  • Garmin intern project

    Conducted technical testing and validation for a large Fetch AMR robot, for use in warehouse automation applications.

  • Garmin intern project

  • Garmin intern project

  • Garmin intern project

  • Genesis Systems intern

    Used Python to create a calculator GUI, which could be used to determine the optimal laser engraver settings for working on curved parts.

Checkout my Resume!

Current as of January, 2025.

Wanna work together?

Interested in getting in touch? Send me an email and I’ll get back to you as soon as I can! :]

Email
wphoward09@gmail.com