Butlr

The coronavirus pandemic has cost tremendous financial lost for retailers and also put their employees' and customers' health at risk. Butlr aims to help retailer safely get back to businesses and increase revenue after the pandemic. 

Butlr protects the health and well-being of retail employees and customers with a low-cost thermal imaging system that captures people’s behavior in indoor and outdoor environments. Butlr’s proprietary hardware pairs with advanced AI to gain insight into people’s movements in physical space, providing actionable intelligence to supermarkets, big-box retailers, homes, and offices, all while safeguarding privacy.

In response to the Coronavirus pandemic, Butlr has implemented features including maximum occupancy control, automated queue and social distance management, and temperature screening. These same features will automatically provide powerful business intelligence to substantially increase revenue for retail chains after the pandemic ends. Butlr has a signed contract to provide these features to 200,000 retail locations worldwide.

For grocery chains, below are the main issues they are facing in the pandemic:

-High exposure risks for staff & customer

-Unknown wait time

-Emergency maximum occupancy regulation

-Overwhelmed staffs during peak time

The problem our grocery chains client have now and many other stores gradually opening up is the increasing COVID-19 exposure risk towards their customers and employees. Customers wait outside based on guesswork. And for the retail operation, it’s extremely risky for 2 or 3 staff members to manually manage the queue and occupancy, Not to mention they are extremely overwhelmed during peak hours. 

Our product reduces the designated labor for crowd control and the exposure risks for staff members.  It automates the process for the staff members with push notification and speaker outside the store.Customer also have a live feed of the queue length to avoid getting into a crowded supermarket.

Apart from grocery stores, most of other retailers suffer even greater lost financially. It is predicted that retail will see a $2.1 trillion loss globally in 2020 due to Coronavirus Pandemic, Butlr can help them get back to business safely and sooner.

Butlr provides a non-invasive technique to localize and model human behavior by marrying advanced sensor technology and state-of-the-art computer algorithms. In light of the current COVID-19, our technology would allow us to monitor and analyze human behavior in physical spaces to contain the spread.

The physical component of the system consists of a low-power and low-cost wireless sensor network system, which was designed and developed in the lab at Berkeley. On-board with a thermal-array sensor, each node gathers 64-pixels temperature data from the environment continuously, detecting disturbances in the temperature profile due to human activities. The networked sensors work in conjunction with computer algorithms to localize humans and classify human behavior in real-time. We trained a neural network to classify human posture and behavior using the data produced in the MIT Media Lab. Compared to the traditional method using intrusive technology such as VGA camera or attaching a beacon to the object, our system is entirely passive, non-intrusive, and stand-alone with additional features such as low-cost and easy to deploy. The novelty identified by our project are: 

• turns any-sized physical space responsive to human activities and behavior. 

• preserves privacy and anonymity. 

• does not require inputs or beacons attached to humans.

We identified two key segments for our technology: retail and assisted living. The common theme is the need to monitor human activities and model human behavior without violating their privacy.

We are working with a US supermarket chain (99 Ranch Market) and installed our sensors at multiple locations. In the current event of COVID-19 when social distancing is a crucial factor in reducing the risk of virus spread, we developed a system to manage the queue in front and inside of their stores as well as automate capacity monitoring. Our technology is useful to both the management at the supermarket as well as the individual consumer. The shoppers can monitor how crowded the store is so they can properly maintain safe social distancing measures. Likewise, store managers is able to use the data to manage crowds and ensure optimal store efficiency.We are also well suited to deploy our system to nursing homes to monitor the activities of the elderly in order to reduce the necessity of having a nurse checking on them all the time, which increases the risk of infection. We are working with a Japanese builder, the Shimizu Group, to develop a POC project on this focus.

butlr' solution for retail sector not only slows the virus from spreading and also help retail owners get back to businesses. in the long run, it could gather data for general traffic and protect customers when similar pandemic happens again.

In assisted living, butlr can:

• Establishing autonomous control to appliances with user behavior sensing and feedbacks
• Generalize resident's activity trends and behavior patterns into health information
• Respond to accidents/unusual activities with real-time notification (e.g. accidental falling).

We designed and developed a system that overcomes all the difficulties existing in the traditional technology.

-VGA camera: In an increasing number of scenarios, VGA camera-based systems are banned due to its risk of privacy violations. Our system does not rely on footages, nor do we collect biometric data of the user, minimizing the privacy concern due to the choice of technology. Our system, due to its wireless nature, can save up to 90% deployment cost when compared to wired camera systems.

-Beacon technology: requires additional hardware that users have to remember to attach to their bodies for the system to function. Our system does not require hardware on the users to achieve those functionalities.

Traditional solutions for activity monitoring are labor-intensive, raise privacy issues, and only provide delayed results. In addition to privacy concerns, there are natural disadvantages of camera solutions to produce real-time results. There is no rocket science behind camera solutions, just stacks of GPUs.

Butlr’s solution generates real-time analytics on user location, trajectory, and shopping behavior. Butlr not only provides instant feedback to help businesses optimize their product placement and customer engagement but also enables personalized shopping experience to maximize their profits.

Butlr’s solution is more cost-efficient and scalable for generating real-time shopper trajectories and behaviors. Technically, we are more capable and adaptable than other solutions. We also help existing solutions solve their blind spot problems. 

Butlr’s Heatic ™ sensor is a passive infrared array sensor capturing body heat. It protects privacy by design. Butlr’s hardware is 90 percent cheaper than a camera and generates results in real-time. The Heatic ™ sensor collects 64 pixels of thermal data per frame, compared to millions of RGB pixels per frame from a camera solution, which dramatically reduces the computation and data storage cost. The Heatic ™ battery runs more than 2.5 years and requires no wiring.

For the past 10 years, Butlr’s team has been working on low cost, low power, non-intrusive sensor networks to reconstruct user behavior.

The physical component of the system consists of a low-power and low-cost wireless sensor network system, which was designed and developed in the lab at Berkeley. On-board with a thermal-array sensor, each node gathers 64-pixels temperature data from the environment continuously, detecting disturbances in the temperature profile due to human activities. The networked sensors work in conjunction with computer algorithms to localize humans and classify human behavior in real-time. We trained a neural network to classify human posture and behavior using the data produced in the MIT Media Lab. Compared to the traditional method using intrusive technology such as VGA camera or attaching a beacon to the object, our system is entirely passive, non-intrusive, and stand-alone with additional features such as low-cost and easy to deploy. 

technology demo:

https://vimeo.com/user82092658/review/398135800/e1403bef36

we signed a contract with a system integrator WiWide Wanda, where they specifically asked to test the accuracy of our product in more than 20 real-world scenarios. we delivered the result and moved to the second stage of our collaboration. 

We've conducted over 130 in person interviews for our sensor applications in retail space. During the pandemic, we have many clients reached out to us asking if we can launch new features to safeguard their employees using the same technology. We are doing pilot projects with multiple companies (many are huge).

500 stores currently

100k stores in one year

800k stores in five years

- Short term: Butlr will continue to focus on helping our retail customers to strive during and after the pandemic by transforming them to become smarter and safer. We are receiving incredible demands on these features and finalizing contracts that will bring our ARR to over $2M. And we will continue to rapidly scale it by leveraging on our channels of system integrators and champion customers.

- Long term: Butlr's vision is "connecting users to the built environment" to bring responsive experience and seamless automation that understands user's need. It means that we make traditional environments like retail, office, and homes context-awared and understand users' needs.

We start from retail to create safe, smart and personalized shopping experience.  Other clients have also paid us to use our technologies to in assisted living, and smart living--where cameras can’t go, to provide analytics on users’ well being and seamless automation. 

Starting with the $60 Billion retail sector as a beachhead market, in the next 10 years, imagine a physical world that truly understands customers, user experience, is safe, responsive, and intelligent. Powered by butlr.

The technical objectives are based on the customer interviews we conducted during the NSF I-Corps Teams program. There are a few recurring demands by customers when considering to acquire a new technical product, which has been converted to the following technical objectives: 

1. The deployment of the hardware components of the system has to be easy and non-disruptive to normal business operations. 

2. The system has to be reliable with high fidelity in a variety of scenarios. 

3. The system needs to produce real-time results.

4. The cost of the system has to be minimized to reduce the capital requirements to acquire the technology.

5. The user interface needs to be intuitive and easy to use.

6. The results of the analysis have to be accurate. 

In addition to the challenges above, we are seeking fundings to help us to fast track the development of applications that are relevant to the current COVID-19 outbreak based on our core technology.

Stemming from those requirements, we carefully reviewed our current design to identify the following objectives and challenges to be overcome with this funding period: 

1.  We are on track to develop an easy to deploy system. Our battery-powered wireless design can minimize the impact on normal business operations during the deployment phase. We are seeking to further reduce the power consumption of the system by optimizing our embedded system design and software implementation. We expect to produce a system that is capable of operating multiple years with a single battery. This feature will further reduce the labor required to maintain the system and reduce the operational cost of the system. 

2. We need to evaluate the impact of partial system failure on the overall effectiveness and reliability of our system. The overall impact needs to be minimized and mitigated to prevent system failure.

3. We are working toward a more light-weight algorithm in anticipation of a large-scale deployment of our system to reduce computational cost. 

4. We are designing a user interface for the user to operate our system without the requirement of domain-specific knowledge of sensor and network technologies. 

5. We are also working to improve the accuracy of our algorithm in determining the trajectory of individuals in large spaces moving across multiple sensors. 

7 fulltime

Butlr is a spin-off company from City Science Group at MIT Media Lab(list of spinoffs). MIT City Sicence Group is a world-leading laboratory for spatial analytics and smart city, with over 15 years' experience on low cost, low power sensor network to reconstruct user behavior without using footages or camera that is intrusive. City Science group won two Most Influential Research Awards of IoT: Ubicomp 10 Year Impact Award ’14 and ’17.

The Butlr. team has extensive knowledge and experience in user behavior reconstruction, ambient sensing network, and multimodal behavior pattern recognition in physical space. We have won multiple awards including Fast Company World Changing Ideas with the technology.

Butlr's technology was granted by the National Science Foundation (NSF) Innovation Corps Program with $50,000 and customer discovery coaching for awarded deep technology. With the help from the mentor from the National Science Foundation. We've conducted over 130 customer interviews to identify the need and refine the business model. We identified omnichannel grocery chains, fashion retailers’ growing needs for precise analytics, contextual promotion, and creating personalized shopping experiences in their existing and new physical locations. This requires a system that can track individuals’ indoor position and understand their behavior in real-time. The current VGA camera systems and other existing solutions can’t fulfill these needs in a scalable way.

- An over 2,000 locations mobile phone retail chain 

- Wanda Group, own 500,000 retail locations worldwide. 

- Century, one of the largest RFID integrators for retail stores. - Regional Grocery chains like 99 Ranch market

Butlr offers two tiers of product through an annual subscription model; pricing is based on covered square footage, and features selected.

Each sensor is $200/yr for basic package, $300/yr for advanced package. To cover a 15k sqft Trader Joe's size store, it requires 40 sensors.

-The basic Butlr package covers foot traffic analytics. This gives the retailer a dashboard like Google Analytics, to better understand customer interest while optimizing product placement and customer engagement. In the pandemic, it also provides maximum occupancy control and automatic queue management.  For retail stores, this reduces the designated labor for crowd control and the exposure risks for staff members. The live queue information is also provided to customers to help them pick the right time to shop, without having to go into a crowded supermarket.

-The advanced Butlr package provides full store real-time trajectory tracking for each individual shopper. Based on each individual’s shopping journey, this creates personalized shopping experiences and contextual promotions for each individual shopper to convert more sales for the stores.

We also license our technology out to retail system integrators.

We are already making profit. Currently, we are reinvesting these renvenue to focus on growth.

We got tremendous tractions from retail and health sector in the Pandemic. With MIT Solve, we want to scale the team and resources so we can fulfill the needs from our clients and scale the company fast.