Category: Yilu’s Status Reports

This week, we focused on testing the fully wireless version of TrashDash as part of our final validation phase. While the system remains functionally integrated, we identified power consumption as a major issue, particularly due to the high energy demand of running ML model inference on the onboard hardware. This currently limits stable untethered operation and is our primary bottleneck moving forward. We are exploring potential optimizations and hardware solutions to address this constraint while maintaining system performance. In parallel, I am taking full responsibility for designing and finalizing the demo poster to clearly present our system, results, and key contributions for the final demonstration.

This week, we made significant progress on system integration and hardware assembly. All components were successfully mounted onto the car chassis, resulting in a fully integrated platform. We implemented and validated obstacle avoidance, enabling the system to safely navigate around objects while moving toward the user. In addition, I developed an emergency stop and resume feature, allowing the system to halt immediately when needed and continue operation smoothly. The motor subsystem was thoroughly tested, confirming reliable movement in multiple directions, including forward, backward, lateral shifts, and rotation. Overall, the system is now functioning as a cohesive unit with core mobility and safety features in place.

During this process, I had to quickly learn several new tools and concepts, especially related to ROS2 integration, hardware-software interfacing on Raspberry Pi, and sensor-based navigation logic. I relied heavily on practical learning strategies such as reading official documentation and debugging through trial and error. I also used tutorial videos and example projects to understand system integration patterns. This hands-on, iterative approach was very effective, allowing me to rapidly identify issues and apply solutions directly to our system.

This week, our team completed the interim demo, where we showcased the integration of the voice, camera, and motor subsystems, demonstrating that the system can respond to voice commands and trigger movement based on camera input. Since the ToF sensors were not yet mounted, we presented the obstacle detection functionality separately using an Arduino setup for better clarity. In addition to the demo, we ordered additional components including a UPS for the Raspberry Pi and camera mounts to securely attach the camera on top of the trash bin for improved stability.

As we move into the verification and validation phase, I will be focusing on the motor subsystem by testing basic motion commands (forward, backward, lateral movement, and rotation) to ensure correct response to control inputs, as well as evaluating consistency and responsiveness under repeated trials. For verification, I plan to measure latency from command to actuation and compare actual vs. expected movement (e.g., direction accuracy and speed consistency) to ensure the motor controller meets our design requirements. For validation, we will test the fully integrated system by evaluating whether the robot can reliably move toward a detected hand within the required time and accuracy constraints in realistic environments. There were no major changes to the system design, but integration and mounting remain potential risks, which we plan to address through early assembly and incremental testing.

Next week, we will focus on mounting all components onto the chassis and begin implementing the follow-human logic as the next key feature.

This week, I assembled the car chassis and researched how to interface it with both the Raspberry Pi and STM32 board. I also set up the development environment and wrote some preliminary motor control code to test basic functionality. In addition, our team attended an ethics lecture, which prompted us to think more carefully about potential issues such as user safety, privacy, and reliability. For example, since our system involves autonomous movement and sensing, we need to ensure it avoids collisions reliably and does not unintentionally collect sensitive data from users or their surroundings.

Overall, there are no major blockers at the moment, but we are slightly tight on time with the interim demo approaching. The car chassis arrived later than expected, which caused a small delay in hardware testing. To stay on track, we plan to increase our workload next week and focus on accelerating integration and testing progress.

This week we made progress on the hardware side of the project. The car chassis and the voice recognition module arrived, which is such a relief as we were worried that it’s not going to arrive on time. I spent time assembling the car chassis platform and bringing up the motors to verify that they work properly. I tested the basic motor functionality and confirmed that the chassis and motors operate as expected, which will allow us to start integrating the control system in the next stage. In addition, I worked on the ethics assignment, which made me reflect on the broader impact of TrashDash, especially in terms of user safety, reliability, and how autonomous devices interact with people in everyday environments. Overall, the project is still on schedule and we are ready to move toward further hardware and software integration.

This week we completed the design presentation and finalized the design report for the project. I contributed to several sections of the report, including the abstract, introduction, motor system implementation, project management, ethical issues, and related work. Writing these sections involved organizing our project goals, explaining the technical design of the motor platform, and documenting how the team plans to manage development and address potential ethical considerations.

In addition, I began writing a preliminary motor control driver for the motor platform so that testing can start immediately once the chassis arrives. Preparing the driver in advance should make it easier to bring up the motors and begin hardware testing.

One concern at the moment is the arrival time of the motor platform. Although it has already been ordered, I have not yet received a notification from the ECE receiving desk confirming that the items have arrived. Once the motor platform is delivered, I plan to proceed with hardware setup and begin testing the motor control system.

This week I submitted request forms for key hardware components, including the car chassis, STM32 board, and voice recognition module. All items are expected to arrive next week (definitely before spring break ends). In preparation, I reviewed implementation documentation and found a detailed online guide covering chassis assembly and motor bring-up procedures. I plan to follow these instructions and begin testing immediately once the chassis is delivered.

I also researched battery options to power the full system and identified several viable candidates that meet our voltage and current requirements. After the chassis arrives and we can verify space and load constraints, I’ll finalize the battery selection.

Additionally, our team divided responsibilities for the design report due next Friday. I will be responsible for writing the abstract and intro,  motor implementation plan, project management, ethical issues, and related work portion. Overall, progress is on schedule, and next week’s focus will be hardware bring-up and initial system testing.

This week I created a BoM for our project. Specifically, I focused on researching motor chassis options for our project to make sure we choose something that meets our load, speed, and size requirements. After comparing different platforms, we decided to use the Hiwonder Large Metal 4WD Vehicle Chassis because it supports around a 5 kg load, has suitable speed for indoor navigation, and is mechanically stable enough for our system. It’s also compatible with Raspberry Pi/stm32 control and supports encoder feedback, which will be helpful for motion control and navigation accuracy. Overall, it seems like a strong fit for both our technical needs and our integration plan.

I also worked on the design presentation slides, mainly focusing on the motor implementation section, test/verification section, and the project management portion. I outlined how we plan to wire and control the motors, what components are needed for integration, and how testing will be staged once the hardware arrives. For the planning section, I helped organize our timeline, task breakdown, and milestones so we have a clearer roadmap for the next few weeks.

Right now we’re on track with our schedule and haven’t run into any major blockers. All required components have been ordered, and once they arrive, we’ll begin testing immediately. The goal for the next phase is to verify motor control, load performance, and basic mobility before integrating sensors and camera. Really looking forward to working on the motors!

This week I worked on the use case requirements, technical challenges, division of labor, and schedule sections in the presentation slides. I practiced the presentation for a few times and then carried out the speech on Wednesday (2/4). Additionally, as soon as the inventory list came out, I discussed with my teammates and reserved the Oak-D Pro Robotics Camera as our main device for depth sensing/hand detection and RPi 5 as our computation unit. I also did some research on the motor platforms that are compatible with RPi5, and found some great options. 

Our current progress aligns with the schedule. All planned milestones for this week, including proposal preparation, presentation, and early hardware planning, were completed as intended. 

In the coming week, I plan on finalizing the motor platform decision and start ordering soon after the meeting on Monday. I will also start preparing for the design presentation slides, working on finalizing the connections/protocols between different components.