On a school trip to Delhi’s Sundar Nursery in 2023, Mahi Malhani was trying to take in the heritage park’s history and restoration. But her attention kept returning to something else: wrappers, bottles and food waste lying on the ground, even when dustbins were close by.
“I noticed litter everywhere, even near dustbins. People often take the easy route and throw waste wherever they can. I wanted to bridge that gap with technology that helps, rather than just tells people to clean up,” she recalls.
India’s waste problem is vast. According to a 2024 Press Information Bureau release, citing the Central Pollution Control Board’s 2021-22 annual report, India generates an average of 1,70,338 tonnes of solid waste every day, of which 91,512 tonnes is treated.
For Mahi (17), now a Class 12 student at Amity International School, Mayur Vihar, Delhi, that problem became the starting point of an idea.
“It was our annual school trip, where we usually visit heritage parks to learn their history,” she explains. “Even when bins were there, waste did not reach them. That made me realise the problem had two parts: littering, and convenience versus responsibility. There had to be a way to bring the two together, and maybe even eliminate the problem.”
From a young age, she had been fascinated with technology. She would tinker with small gadgets, dismantle old electronics, and experiment with coding. “I started learning Python, then explored C++ and JavaScript,” she tells The Better India.
“I wanted to understand how machines think and move. If I can teach them to act intelligently, like humans, but more precisely, they can help solve everyday problems, like managing waste efficiently.”
How I went from sketches to a prototype
Motivated by this thought, Mahi began exploring possible solutions. She started sketching ideas and researching prototypes that could automate waste collection and sorting, combining convenience with responsibility.
Mahi’s idea became TRASHbot, a robot that can collect waste and sort it on its own.
In August and September 2023, she dived deep into the challenges, noting that any effective solution would need to be intelligent, autonomous, and user-friendly.
Her ideas soon needed guidance to move from concept to reality. That is when she connected with Rancho Labs, a robotics and IoT initiative at IIT Delhi. During a summer programme in August 2024, she visited the IIT campus for a workshop and spotted a Rancho Labs poster. Curious, she approached the team and, with her parents’ support, conducted a survey and met the professionals. “They helped and guided us, explaining practical aspects of robotics and IoT,” she says.
Her mentors, including teachers from Rancho Labs, and her school computer teacher, Mrs Deepshikha Sethi, provided critical technical guidance while validating her ideas. “They were always available to answer questions and suggest improvements,” the 17-year-old explains. “But the prototype itself was my own creation, and that independence was very motivating.”
A robot that sees, moves, and sorts
Mahi’s idea became TRASHbot, a robot that can collect waste and sort it on its own. She did not want to build a dustbin that simply moved around. She wanted it to notice waste, move towards it, sort it correctly, and make people curious enough to use it. “I wanted it to feel alive, almost like a companion in cleanliness,” she says.
TRASHbot does two main things. It moves around safely, and it separates waste into the right compartment.
To move around, the robot can either be controlled through a smartphone app or left to roam on its own. It uses ultrasonic sensors, which work a bit like the parking sensors in cars. They send out sound waves, sense nearby objects, and help the robot avoid bumping into people, tables, walls, or anything in its way.
“It is aware of obstacles, just like we are. If it encounters something in its path, it either stops or manoeuvres around it,” she explains.
For sorting waste, TRASHbot uses a camera and an AI model called YOLOv3. When it sees an item, it takes a picture and checks it against images already stored in its system. This helps it decide whether the waste is biodegradable, like food scraps, or non-biodegradable, like plastic wrappers.
Once it recognises the item, small motors tilt the correct compartment so the waste falls into the right section. “It sees, recognises, and sorts, just like a human, but faster, and without mistakes from carelessness,” says the young innovator.
Inside the robot are two small computers that divide the work. The Raspberry Pi works like the robot’s brain. It studies the images captured by the camera and decides what kind of waste it is. The Arduino Uno works more like the robot’s muscles. It controls the movement, helping the robot turn, stop, or move ahead.
Six small motors help TRASHbot roll across the floor, while two turning motors tilt the waste into the right compartment. A lightweight body keeps the robot balanced and easy to move. “It may sound complicated, but for the user, it is as simple as pressing a button or letting it work on its own,” she says.
From home trials to school cafeterias and societies
TRASHbot was first tested at Mahi’s home, then in her school cafeteria during busy lunch breaks. “Seeing it move between tables and collect wrappers was incredible,” says Dr SK Singhal, Physics HOD at Amity International School. “It engages students with its LED lights and sensors. You feel part of the process without lifting a finger.”
To move around, the robot can either be controlled through a smartphone app or left to roam on its own.
The next step was testing the device in residential societies. Gita Nath, joint secretary of Lotus Boulevard Society, recalls her first encounter, “The robot sorts waste automatically. You just throw it in, and it tilts the compartment to the right place. It is convenient, innovative, and makes day-to-day waste management effortless.”
Yash, a 17-year-old resident of Lotus Boulevard, shares, “During events with food stalls, the device was extremely helpful. It automatically sorted all waste, making it easier for people to dispose of it responsibly.”
When cleanliness became as simple as opening an app
Since March 2025, TRASHbot has been put to work in two residential societies in Noida and with a municipal corporation in Udaipur. Around 15 to 20 people have used it directly so far, and for Mahi, each interaction matters. Every time someone watches the robot approach waste and sort it correctly, it becomes a small lesson in keeping surroundings clean.
And the robot is holding up well. It correctly identifies waste about 85% of the time, sorts it into the right compartment in 90% of cases, and avoids obstacles with a 98% success rate. On a full charge, it runs for around two and a half hours, enough to make several rounds before it needs to be plugged in again.
Using TRASHbot is intentionally simple. Through the app, users can guide it forward, backward, left, or right in manual mode. Or they can switch to automatic mode and let it move around on its own, encouraging people to throw waste into it without having to walk over to a dustbin.
“I wanted anyone, even someone using it for the first time, to feel comfortable with it within minutes,” says Mahi.
What it took to make the robot work in real life
The journey to TRASHbot came with several setbacks. In early versions, voltage drops caused the Raspberry Pi to reboot. At other times, delays in Bluetooth communication meant the robot took longer to identify and sort waste.
Cost was another concern. Building the first version required nearly Rs 20,000. “After refinements, the cost came down to Rs 7,000, with the potential to reduce to Rs 4,000 using higher-quality components,” Mahi says.
Then came the real test: taking TRASHbot out of controlled environments and into everyday spaces. Uneven surfaces, rain, crowded areas, and constant movement around it required changes to its sensors and mobility.
For Mahi, every obstacle became a problem she had to solve. “Every time the robot faced a problem, I had to find a solution that made it strong, safe, and efficient,” she recalls.
What started as a school project soon earned recognition beyond the classroom. In 2025, Mahi participated in the World Robot Olympiad, securing first rank in Delhi and second rank in India in the virtual championship.
She was also a finalist at the Inter-Amity Science Fair, ‘Vasudha’, and TRASHbot was selected among the top 50 AI-based projects from over 3,500 entries across the country at the India AI Impact Summit.
For sorting waste, TRASHbot uses a camera and an AI model called YOLOv3.
“The questions from judges helped me refine TRASHbot and think critically about real-world applications. Participating in competitions also offered her exposure to other young innovators, inspiring improvements in the device’s design and functionality,” she says.
For residents, TRASHbot has changed daily habits. Gita shares, “We no longer need to sort waste manually. The device automatically separates it. It is convenient and has improved cleanliness in our society.”
Dr Singhal adds, “In school, the device confirms cafeterias remain clean, even during busy lunch hours. Students learn recycling, AI, and ethics in action without realising it.”
Yash emphasises the behavioural impact, “It is teaching people habits along with collecting trash. Students and residents are more conscious now because the robot makes cleanliness visible and interactive.”
The device’s design promotes engagement. LED indicators alert users when it is ready to collect waste, while its movement attracts attention and encourages participation. “People respond positively when they see it moving around,” says Mahi. “It is almost like it’s inviting them to be part of the process.”
Looking to the future
The young innovator has ambitious plans for her device. Currently, it classifies waste into two categories—biodegradable and non-biodegradable. She aims to expand this to five categories: plastic, metal, glass, biodegradable, and non-biodegradable.
Other improvements include Wi-Fi connectivity for cloud-based monitoring, allowing schools and societies to track waste collection remotely. Solar panels could extend battery life, and the app could be adapted for larger public spaces. “Rural areas present unique challenges, with rougher terrain and fewer facilities,” she explains. “The goal is to make TRASHbot effective everywhere, not just in cities.”
For Mahi, TRASHbot began with one question: why does waste still end up on the ground when dustbins are close by?
TRASHbot was first tested at Mahi’s home, then in her school cafeteria during busy lunch breaks.
Today, her answer is a robot that moves towards the problem, makes disposal easier, and reminds people that cleanliness often begins with a small nudge. Since its first pilot in March, TRASHbot has shown what can happen when a teenager looks at an everyday civic problem and decides to build a solution herself.
All pictures courtesy Mahi Malhani.
