Blogs

4. Does Simplicity Scale? Reproducing and Stress-Testing SimpleView for Point Cloud Classification

4. Does Simplicity Scale? Reproducing and Stress-Testing SimpleView for Point Cloud Classification

Can a simple 2D projection model really match complex 3D architectures on point cloud classification? We reproduce SimpleView, then push it onto new datasets, hyperparameter sweeps, inference optimizations and ablations.
3. Reading the Label, Not the Object: A Typographic-Distractor Control Dataset for CLIP-Based Robot Grasping

3. Reading the Label, Not the Object: A Typographic-Distractor Control Dataset for CLIP-Based Robot Grasping

Does CLIP ground a robot’s language query in what an object looks like, or in text written in the scene? A controlled 400-image dataset built from real OCID-VLG scenes, and what it reveals about typographic attacks on CLIP-based grasping.
2. Reproducing GeoLanG: Geometry-Aware Language-Guided Grasping with Unified RGB-D Multimodal Learning

2. Reproducing GeoLanG: Geometry-Aware Language-Guided Grasping with Unified RGB-D Multimodal Learning

A from-scratch reproduction of GeoLanG, a model that uses language and depth to decide where a robot should grasp an object, plus ablations showing which of its modules actually help.
1. Delta Cup 2021 - A motivation to display an engineer's skills

1. Delta Cup 2021 - A motivation to display an engineer's skills

Before checking out this post, please look up the Projects section to get an overview of my work. Why do you think I describe the Delta Cup as a test of an engineer’s skills? It is because, in my opinion, practical knowledge is a key asset for a person to be recognized as an engineer. But don’t get me wrong, I am not against theoretical knowledge. Theoretical learning is the foundation of any concept, but whether it can be applied in practice is a different question.

1. Delta Cup 2021 - A motivation to display an engineer's skills

Planted December 19, 2021

1. Delta Cup 2021 - A motivation to display an engineer's skills

Before checking out this post, please look up the Projects section to get an overview of my work. Why do you think I describe the Delta Cup as a test of an engineer’s skills? It is because, in my opinion, practical knowledge is a key asset for a person to be recognized as an engineer. But don’t get me wrong, I am not against theoretical knowledge. Theoretical learning is the foundation of any concept, but whether it can be applied in practice is a different question.

Delta Cup is a competition organized by Delta Electronics. The contest is hosted by the Chinese Association of Automation (CAA) and sponsored by the Chinese Association of Automation and the Industry and Information Industry Vocational Education Teaching Steering Committee under the Ministry of Education. Every year the competition has different tracks, and this time they were “Innovative Machines”, “Smart Factory”, and “Better Future Living”.

It all began with a slight push from our seniors at SRA (Society of Robotics and Automation) to participate in the competition. Dhairya Shah, Ninad Jangle, and I formed a team. Initially, we brainstormed ideas. Several were rejected because of their drawbacks on the practical side. We finalized our idea of a pick-and-place bot under the “Smart Factory” track, built a static model in simulation, and made a rough list of Delta products that would be apt for the model. Mr. Alister D’Silva, from the Delta side, was mentoring teams from the Mumbai division. He reviewed our idea and gave us his feedback and advice. We implemented the changes and described the problem and our solution in the proposal.

In May, the results were out. We were startled to hear that we had been selected for the international round, competing with 100 other teams, and were one of the two teams representing the Indian division!! So, our immediate next step was to finalize the list of Delta products and distribute the work. Dhairya and Ninad handled the 3D bin packing algorithm for efficient packing and the automation of the bot, controlling it using a PLC and HMI. I took charge of the mechanical design and fabrication of the bot.

As we had only 2 months, the workload was immense. Building up the structure was our main challenge as we were naive in that field. Prof. Faruk Kazi was our mentor. He assured us funding from our college for the setup and assisted us in solving problems from the college side. Through him, we met Dr. Sangram Patil, who was himself a Ph.D. graduate of VJTI. He had a mechanical background, and his mentorship and support helped us build the bot even faster. He helped us simplify the design, and his one piece of advice was “Due to the time constraint, go with standard structures while designing and choose a material which has a good supply and which has the capability of handling loads”. He connected us with his friend Mr. Amit Ajmera, who came to our rescue. He introduced us to various types of machines and manufacturing problems. I got the opportunity to view the laser cutting machine and hear about its working. And under his supervision, all the work was done.

After a lot of hard work and many sleepless nights, we were finally able to complete the design and successfully manufacture the bot. Though we weren’t able to automate it due to the time crunch, we presented a polished product to the judges. We won the 2nd prize and were overwhelmed by the whole experience. I am never going to forget these 3 months of my life, as they taught me that it’s easy to imagine but it’s a lot harder bringing it to reality.

Problems Faced

  • In April and May, COVID-19 cases were at their peak here in Mumbai, so we had to take risks and go out to get our work done.
  • Industrial production rates were significantly reduced, so there were delays in the transportation of parts.
  • Coordinating with people and explaining the process of implementation.

Our Experience and Learnings

  • Software and hardware, the technicalities of industrial robots, and automation equipment like PLCs and HMIs
  • Mechanical production and manufacturing in industry
  • Time, money and resource management
  • Team bonding and building connections
  • Practical work is time-consuming, and we experienced real-world problems firsthand

Special thanks to: VJTI, COE-CNDS VJTI, SRA VJTI, Omkar Bhilare, Saad Hashmi