Building AI-powered aqua drones that collect waste, weed, and data from waterways
Pitch us on your solution
Problem: Keeping urban waterways clean is essential for healthy cities and also preventing ocean plastic pollution. However, the current global methods for waterway maintenance are labour intensive, inefficient and don't scale.
Solution: We’re automating waterway maintenance with AI and drone technology. We are building all-electric, autonomous aquadrones that collect floating debris and remove weed using computer vision. They also function as mobile data acquisition platforms, tracking waste aggregation trends while recording water and air quality data.
Impact in Bangladesh: Cities in Bangladesh are rapidly expanding. Consequently, urban waterways Bangladeshi cities are being destroyed by solid waste dumping, unauthorized land filling and construction, etc. Our drones make waterway cleaning cheaper, more effective and more efficient. Further, data acquisition helps measure chemical discharge, intensity of pollution, canal depth, water carrying capacity etc. informing sound intervention by authorities.
Film your elevator pitch
What is the problem you are solving?
Dhaka has lost up to 50 canals in the last 30 years - due to severe solid waste pollution and encroachment. This has led to the following issues: 'a collapse of the natural drainage system leading to urban flooding in normal monsoon, spread of skin and intestinal diseases among the urban poor who use canal water in the absence of access to clean water, breeding of mosquitoes leading to water borne diseases like malaria and diarrhea in stagnant canals, as well as environmental impacts such as loss of landscape, decreasing water quality, reduced ecological connectivity, and decline of ground water aquifer recharge'' (Environmental Justice Atlas 2015). This is not just the story of Dhaka, this is the story of nearly every Bangladeshi city.
Dhaka is home to 18 million people. They are one of the worst affected urban populations in the world when it comes to flooding. Korail slum in Dhaka spreads over 50 acres and houses more than 50,000 people - and during monsoon the entire slum can be found underwater. Bangladesh is home to 50 million urban residents of whom about 20 percent live below poverty line.
Contributing factors: population rise, unplanned urban development.
Who are you serving?
We serve two vulnerable segments of the population:
- Slum communities in densely populated urban areas, particularly those lining waterways
- Coastal communities dependent on the marine ecosystem for livelihood and food
Urban waterways in developing countries are narrow, shallow, and heavily polluted. To perform cleaning in such conditions, the device needs to be compact but capable of heavy-duty operation and offer a large payload. Further, the pricing must be suitable for developing countries. Our trash skimmer encompasses all these considerations and will have 10x the carrying capacity of comparable solutions in this space.
Further, we have observed and interviewed a large number of cleaners, who also live in the same slums. Our interaction with the cleaners led us to decide that we will primarily sell human operated versions of our trash skimmers in regions where there is a shortage of jobs. Thus, we will not cause a loss of employment: some of the existing cleaners will be retrained as operators while others will be absorbed by the ecosystem assigned to safely dispose of the collected waste.
Our solution will ensure that the waterways close to slums are clean, flowing, and free of larval breeding locations.
What is your solution?
Our primary product is an intelligent trash skimming aqua drone that uses computer vision to scavenge for floating waste and aquatic weed. Currently, we finished the design and engineering work and started manufacturing the product. We expect to complete the work in 4 months. We launch as a human operated drone in South Asian market for reasons listed under the previous question.
Key design features:
- All-electric and highly energy efficient
- Superior hydro dynamic performance
- A novel conveyor manipulation system for independent collection and disposal of trash
- Small form factor and lightweight
- Dimensions: 3 m (length) x 2 m (width) x 0.75 m (height)
- No-load weight: 400 kg
- Average speed: 10 km/h
- Run time: 8 hours
- Payload volume: 1,000 liters
- Payload weight: 200 kg
- On-board sensors: RGB camera, GPS, underwater sonar, depth, temperature and pH-level sensors
- On-board computer: Jetson TX2
In parallel, we are working on the robotics software platform using a smaller, proof of concept drone we have built.
The test drone's current workflow is:
- An operator defines a target search area with GPS points
- The aqua drone then scavenges for floating waste using computer vision within that geo-fence. We use k-means clustering and Speeded Up Robust Features for identifying floating debris and optical flow based tracking to continuously track and collect the waste
- The drone returns to the starting position once it completes the work.
- The operator can view the on-board camera's live stream and also intervene at any point during the mission via a base station (either mobile phone or a laptop) that is linked with the aqua drone.
Notes: The communication is via 3G/4G (any distance, provided both the robot and the base station have strong GSM connections) or WiFi (100 meter radius with line of sight). Additionally, the drone also supports Radio Control (1000 meter radius with line of sight). The test drone is currently trialing alongside Colombo's canal cleaners.
Select only the most relevant.
Where our solution team is headquartered or located:Colombo, Sri Lanka
In which sector would you categorize your solution?
Our solution's stage of development:Prototype
Describe what makes your solution innovative.
We are innovating on two fronts.
On the hardware design front, when compared to other available mechanized solutions in this domain, our drone (i) is all-electric and highly energy efficient, (ii) of a smaller form factor and lightweight - making it amenable for a variety of use cases and easy transportation, (iii) can collect and safely dispose trash without additional infrastructure or support, using a novel single conveyor manipulation system, (iv) has built-in fail safes, (v) measures waste quantity, water and air quality, waterway depth, and comes (vi) geared for complete automation.
On the software front, we have already built an machine learning platform that enables our existing proof of concept to autonomously detect, localize, and collect floating trash using an on-board camera; coupled with underwater sonar and GPS for safe positioning and navigation. Initial product we launch in South Asia will be human operated, due to socio-economic, safety, technology maturity, and regulatory concerns.
Fully autonomous system uses computer vision and machine learning in the following ways:
(i) segment floating waste and weed from water surfaces, make decisions on what to collect and avoid.
(ii) 3D localize the trash to be collected using structure from motion.
(iii) infer absolute size and depth of debris by fusing IMU measurements.
(iv) probabilistic positioning (SLAM) using a combination of vision, GPS, and underwater sonar readings.
(v) learn over time where most waste typically aggregates and pre-plan cleaning mission accordingly.
(vi) continuous determination of best possible navigational course to extend battery life.
Why do you expect your solution to address the problem?
Problem: in developing countries, urban poor disproportionately live next to polluted inland waterways. This places them under severe risk from spread of vector borne disease and other illnesses such as diarrhea.
Our drones perform the following activities: (1) effective and regular cleaning of waterways; (2) continuous monitoring of water quality parameters (pH level, total dissolved solids, etc.); (3) identifying and disturbing larval breeding grounds (feature to be built).
Direct outcomes of our activities are waterways with vastly reduced floating and semi submerged debris, less stagnant waters, and deep information and understanding on the quantity of waste, their sources and trends, location of larval breeding grounds, and health of urban water bodies.
The key short-term outcomes are cleaner waterways next to slum communities, reduction in vector breeding grounds, and execution of interventions that stop pollutants at source, progressive reduction in the amount debris outflow into the ocean.
The key long-term outcomes are significant reduction in the incidence of vector borne diseases among urban poor population in developing countries.
Select the key characteristics of the population in Bangladesh your solution serves.
In which countries do you currently operate?
In which countries will you be operating within the next year?
How many people are you currently serving with your solution? How many will you be serving in one year? How about in five years?
Currently, we are trialing our existing prototype in a 1 km stretch in Colombo North. As such we are currently enabling better surroundings for roughly a thousand people or 200 families.
In the next 1-1.5 years, we will be operational throughout Colombo Metropolitan Area: our drones will cover the city's entire canal network, cleaning over a 1 million sq meters of waterway, stretching 100 kms. The total slum population adjacent to the canals is close to a million people. They will experience cleaner waters and significantly reduced risk from vector borne disease and floods.
In the next 2-3 years, we will be operational in waterways across 10 major cities in Bangladesh and India. Our sales target over this period is to put least 300 drones to work in our urban waterways in the first 3 years. We estimate cleaner waterways in these cities will directly impact 50 million people from slum communities.
In addition, we will be stopping an estimated 10 million kilograms of trash, mostly plastics, from entering the ocean just in Colombo upon complete launch. We will be directly preventing and reducing marine plastic pollution and indirectly protecting the lives and livelihoods of coastal communities.
What are your goals within the next year and within the next five years?
Our vision is to tangibly aid at least 200 million urban poor in developing countries, lead lives that are enriched by cleaner waterways, at reduced risk from the spread of vector borne diseases, and climate crisis impacts such as flooding by 2025.
We achieve this by doing two things in parallel. One, opeartionalise our waste removing aqua drones. Two, add to the technology stack of the aquadrone platform.
This is our waste removal aqua drone launch plan for the next five years:
Year 1 - Jan 2019 - Dec 2019: Design and build winning hardware for electric trash skimming aqua drone in urban waterways.
Year 2 - Jan 2020 - Dec 2020: Launch aqua drone across Colombo's canals.
Year 3 - Jan 2021 - Dec 2021: Launch aqua drone in 10 major cities in Bangladesh and India.
Year 4 - Jan 2022 - Dec 2022: Launch aqua drone in additional 20 cities, including in Africa.
Year 5 - Jan 2023 - Dec 2024: Launch aqua drone in additional 40 cities, including in Latin America.
This is our technology road map for the next five years:
1.Exploring additional use cases for the aqua drone - search and rescue during flooding, emergency communications network, applications in aqua culture. (difficulty - medium)
2. Aerial waste mapping - i.e. flying drones that can identify and map waste in waterways, alert authorities. (difficulty - medium)
3. Aerial vector surveillance - identifying mosquito breeding grounds using flying drones. (difficulty - hard)
What are the barriers that currently exist for you to accomplish your goals for the next year and for the next five years?
1. Financial: ability to raise Series A funding in Sri Lanka within the next 2 years. The local startup environment has low liquidity. As an early stage deep tech, Kayal has a high capex requirement. It is difficult to source the depth of funding required for deep tech within the Sri Lankan startup ecosystem.
2. HR: A key barrier is hiring top quality robotics talent.
3. Legal: We will file for a patent once the current design iteration process is over, guarding our unique IP. This is a barrier we will overcome in the next year.
4. Market: The problem of polluted waterways is most acutely felt in developing countries in South Asia, Southeast Asia and Sub-Saharan Africa. Kayal is launching in Sri Lanka and we plan to then expand into markets such as India, Bangladesh, Indonesia, and South Africa within the next three years. Each new market has a unique local business and regulatory environment.
5. The main buyers for our products are central and local governments, ports and other government agencies with maintaining large water bodies. Governments are buyers with a long lead time; however, they can be sticky once the relationship is established.
How are you planning to overcome these barriers?
1. We are applying to international sources for both venture capital and grants. We are also constantly tapping into and growing our external network through startup competitions and cold emails to engage and connect with potential new funding sources.
2. We are developing an HR plan that involves connecting with local graduates who will have the necessary skillset that we are looking for.
3. We are working in stealth mode until we are able to file the IP. Additionally, we will work with the government's institution that files patents to minimize delays in the patent filing process.
4. By applying to international competitions, we are expanding the base of individuals we connect with. We are strategically networking to find people who can facilitate introductions in the markets we plan to expand into. Currently ICTA and ClimateLaunchPad (detailed in sections below) have committed to connecting us with customers locally, in India and other South East Asian countries.
5. We are working to bring relevant politicians or Ministers onto our board to champion the startup and support procurement processes. Recent 2019 budget initiatives encouraging government to favour startup procurement and increase exports is also encouraging Kayal’s position.
If you selected “I am planning to expand my solution to Bangladesh,” please provide an overview of your expansion plans. What is the market opportunity for your business or product in Bangladesh?
Our primary goal through TigerIT partnership is acquiring a pilot with Bangladesh Water Development Board to do a series of trials across Dhaka. We would ideally like to get started after June 2020. By this point, we would have proofed our concept in Colombo with the waterway authority here. In 2021, we plan to launch commercially in Dhaka and expand to cover Barisal the following year.
Dhaka is spending $ 10 million or more annually on cleaning the canals. Barisal, the city of canals, has a $ 70 million project running through 2030 to revive the canal system inside the city. We believe the total urban waterway cleaning market in Bangladesh to be over $ 100 million. This calculatioon does not include cleaning of rivers and other open waterways outside of major cities.
Another immediate commercial opportunity for us in Bangladesh is its growing aqua culture industry which now exceeds $ 50 billion in exports. We are already in talks with the largest shrimp exporter in Sri Lanka for regular weed and algae removal in their reservoirs. Our device is perfect for this application and comes with the additional advantage of zero pollution of aquatic resources. Through TigerIT we would be delighted to reach out to aqua culture companies in Bangladesh through this programme.
Select an option below:For-profit
If you selected Other for the organization question, please explain here.
How many people work on your solution team?
- Computer vision engineer (BSc in Electrical and Electronic Engineering)
- Electronic engineer (BSc in Electrical and Electronic Engineering)
- Industrial designer (MSc in Industrial Design)
- Overall technology adviser (PhD in robotics) - 20 hours a week
- Sales and finance (MBA, Wharton, incoming) - 10 hours a week
For how many years have you been working on your solution?
Full-time for the past 6 months
Why are you and your team best-placed to deliver this solution?
Currently, we have a core team of five people - 3 engineers (1 with a PhD in robotics)*, 1 industrial designer, and 1 business and finance person. Between us we have the following technical skills: electrical engineering; computer vision; robotics; software development; industrial design. We have a team that is conversant in 5 languages: English, Tamil, Sinhalese, Dutch, Tamil (all very strong), and Hindi (weak).
We have spent more than a year studying and understanding the problem domain we’re working on.
We have built a solid network of promoters in India and Sri Lanka who will help us take the technology into governments. We have a partnership with the Sri Lanka Land Reclamation and Drainage Corporation and they are ready to pilot our technology as soon as it is ready.
The founders of Kayal Technologies, Elijah and Janagan, started working full-time on this project in January 2019. At the time, we had not raised funds, and only had a partially evolved idea. Today, we have raised $42,000, built a working prototype, established a partnership for a pilot, and in the process of reaching out to more end users. We are growing fast and are committed to making our technology the go-to waterway maintenance tool across the developing world.
Note*: Dr. Senthan Mathavan functions as our overall technology advisor. He currently spends 20 hours a week at Kayal. He will join full time as CTO starting from January 2020.
With what organizations are you currently partnering, if any? How are you working with them?
Sri Lanka Land Reclamation and Drainage Corporation
- This is the government agency in-charge of waterway maintenance in Colombo. We are currently trialing our prototype with their manual cleaners as we continue to improve our computer vision technology. This institution has also agreed to pilot our first market-ready product in December.
ClimateLaunchpad - Climate-KIC (EU) & ClimateStudio (India)
- We are the national winners of ClimateLaunchpad in Sri Lanka. We are currently going through their accelerator programme which is helping us discover customers in India.
Information Communication Technology Agency - Sri Lanka
- We got funded on 17 July 2019 by the ICTA Sri Lanka, the government institution for IT, through their grant programme for the top 12 startups in the country. We are getting business mentoring now. They will also link us to foreign customers as their 6 month incubator programme progresses.
What is your business model?
We have a three-pronged business model. (i) We sell aqua drones to public and private entities involved in regular waterway maintenance at $20,000 per unit. Payment is received before the machine is shipped. (ii) We plan to charge a subscription fee of $200 per annually from customers who use our fleet management and data visualization web application. (iii) We contract out waterway cleaning services to those entities who require periodic waterway cleaning at $40 per hour (or per 10,000 sq. meters). Here, payment is received upon the completion of the mission.
We estimated that it will cost roughly $7,000-$8,000 to manufacture and sell one drone. Similarly, it will cost $15 per hour (in Sri Lanka) to offer the cleaning service, including electricity for charging, payment for the operator, and transport.
We will employ direct sales for exports to Europe and US markets and the mission as a service model locally in Sri Lanka and in India through a partner.
What is your path to financial sustainability?
We estimate our first year expenses to be around $58,400. We have raised $50,500 of this amount with another $ 8,800 expected before December 2019. In summary, we have a runway of 12 months. We are drawing low salaries and are running a frugal operation - we may be able to stretch the runway to 18 months.
We will require another $3 million coming into the company at the end of the runway to cover the next 30 months. We start making revenue in 12 months from now (July 2020) and will break even in another 26 months (October 2022). This is so if we spend heavily on tooling in 2021 and ramp up our production throughput by 10x from 2020, and then aggressively pursue a 2x year-on-year revenue growth in revenue in successive years.
Our path to financial sustainability is: build market ready product by December 2019, pilot with established partner in Sri Lanka in January and February 2020, source partly paid purchase orders during this period, raise $3 million using the purchase orders as evidence for traction.
We have already set up a Delaware C Corporation in the USA to enable easier fundraising.
Annualized estimates are as follows:
Why are you applying to the Tiger Challenge?
We are applying to TigerIT Challenge for the following reasons:
Receive mentoring and support on product and technology contextualization for the Bangladeshi market.
Access to Bangladeshi market - government actors in water related activities, Bangladesh Water Development Board, private actors in this space, Bangladeshi aqua culture industry.
Connect to Bangladeshi investors in the sustainability / clean-tech domain to help with further fundraising.
Funding to expand our product and service offering in Bangladesh.
Attract top Bangladeshi technology talent to come on board to work for Kayal.
What types of connections and partnerships would be most catalytic for your solution?
If you selected Other, please explain here.
With what organizations would you like to partner, and how would you like to partner with them?
- Bangladesh Water Development Board
- Dhaka Mayor's Office/Dhaka North-Dhaka South City Corporation(s)
- Barisal Mayor's Office/Barisal City Corporation
- Government authorities/private enterprises/ NGOs tasked with waterway maintenance or working in this space
- private and public companies engaged in aqua culture
We would like to engage such organisations for carrying out pilots across major Bangladeshi cities.