Weeks 5-6 (S3)
Hello everyone! We are sorry for the long delay since the last biweekly update. Our writer has been unable to fufill the responsibilities of the role and has resigned. We are looking for a new writer but Executive Lead, Harrison Canning, will take over writing in the interim. This update will cover the he period of 9/29 - 10/12. Next week we will output an update covering 10/13-10/27.
We have been accepted into the 2019 Rochester Maker Faire! This will take place on November 23rd, 2019! To prepare for the event, NXT will be hosting its first hackathon the first weekend of November.
Below I have included a description of what we will be bringing to Maker Faire broken down by team as written by the Project Leads:
We will be presenting an accelerometer-based wheelchair control solution viable for any individual with motor control from the neck up. This will include an initial version of our joystick manipulator device, a head-mounted accelerometer, and electrodes to detect eye blinking and jaw clenches.
The thought keyboard team would like to present an individual using a prototype of the Thought Keyboard. If that cannot be done, a mock-up video will be presented.
We will have demo boards, a poster, and a demo that will have an emg controller, either flex meter or a remote control robot
For Maker Faire the prosthetics team is planning on getting one of the InMoov arms up and running to show off a basic control mechanism that visitors could potentially try as well. We're also planning on bringing a monitor to have the 3D model files of the redesigned hand/arm on display.
Updates by Team
The team has been focusing on the headgear design. They have determined placement, electrode material, and are working to create schematics for the circuitry and designs for the overall headset. They have been familiarizing themselves with the documentation for the ADS 1299s, the main chip we will use on the boards.
Rebuilt our shader system: We used to write each shader four times, requiring us to write 40 shaders for all the platforms that the game runs on. Now we only need to write each shader once in HLSL, and our shader system converts it to other platforms automatically. Now, we only need to write 10 shaders.
Working on a Deferred Lighting System, which will let us to render lights more optimally
We are working on converting our dynamic command buffers to static, which will improve performance. It is also a prerequisite before we start porting the game to Vulkan 4. We are building tools that let us edit the game without having to manually change the code (light editors, A-Star node editors, etc).
Our A-Star algorithm for AI cars works flawlessly when the game first boots, but becomes unstable when used several times. After we fix that, we will have races.
While working with the Keyboard Team, we are looking for the most optimal way to create an OpenBCI Cyton driver. We tried to rewrite the open source Cyton USB driver by OpenBCI, in Windows, with the MinGW compiler. However, it was not compatible with our Visual Studio projects, so we had to rewrite it again. We tried using networking sockets with the OpenBCI WiFi Shield, but we could not get the devices to connect to our WiFi network, so we couldn’t do that. Currently we are looking into LibLSL for the Keyboard Team, and possibly porting LibLSL to game consoles.
We added some new members to the team: two U of R students who will work on machine learning, and one RIT mechanical engineer who will work on a joystick manipulator
During the past week, we got together for a few hours to re-create the setup of the McGill team's Milo wheelchair. The goal was to measure performance, and it was a good learning experience for using wet electrodes and parts of the software. Unfortunately, we ran into some issues with their code; we are reaching out with questions. The mechanical engineer was briefed on the goal for joystick manipulation
In the coming week, we hope to make another attempt at measuring the performance of the MIlo system.
Began to print out parts for Inmoov arm.
Have had poor success rate using the Prusa Multi-material, so we're tweaking belt tensions of extruding head, x-axis and y-axis required.
Organized design specifications for sub-teams to have a proper Idea of what to do.
Thought Keyboard Team
Created Prototype mock ups
Setup a Github for the whole team to use
Split up team into subteams
Machine learning team -- taking online courses in machine learning
UI team should have a working prototype done by 10/27
Human Subject Research
We have submitted our human subject research documents to RIT's internal research board! We will be able to start advertising and conducting the study soon as it is approved! In the mean time, we are looking at getting grant funding and start training the research administrators on how to go through the procedure and use the custom software.