College - Author 1

College of Engineering

Department - Author 1

Electrical Engineering Department

Degree Name - Author 1

BS in Electrical Engineering



Primary Advisor

Benjamin Hawkins, College of Engineering, Electrical Engineering Department


In 1975, four percent of children aged five to nineteen were categorized as overweight or obese. As of 2016, this figure climbed above 18 percent [1]. Researchers at California Polytechnic State University San Luis Obispo (Cal Poly) want to investigate the effect of overeating in early childhood on later childhood obesity. This research requires collecting feeding pattern data on infants, which proves challenging. Parents cannot be relied on to regularly collect clean data due to factors including work schedule, multitasking, and general exhaustion. Thus, we have developed a tool to automatically collect data on feeding frequency and duration, as well as bottle angle during feeding.

The Smart Bottle is an attachment for baby bottles made of a flexible 3D-printed sleeve that slips on to the bottom of the bottle. The attachment itself is a two-inch (5cm) tall plastic housing with a diameter of seven centimeters in diameter to fit typical baby bottles. To collect data, this attachment houses a 3-dimensional accelerometer, a 3-dimensional gyroscope, and a real-time clock (RTC), and a volume sensor (left vague to protect intellectual property). For data storage and transmission, it includes an SD card reader, and a Bluetooth Low Energy (BLE) module. Finally, for power, it has a Lithium-Polymer (LiPo) battery, and a coin-cell battery. Using a neural network, measurement instruments identify feeding events and measure an infant’s feeding duration, feeding angle, and amount eaten. All measurement data is then recorded on the SD card and broadcast to researchers or parents via Bluetooth. The added Bluetooth functionality will increase the drain on the LiPo battery but using a BLE-capable module minimizes this.

After manufacturing and assembly, testing found that necessary systems were all functional. The accelerometer returned numbers that changed with movement, which allows the neural network to detect feeding events. The volume sensor was accurate within 0.5ml as tested with volumes ranging from 10 to 85 ml. The SD card correctly stored data without issue. The BLE module accurately transmitted data to connected devices. The RTC did not lose a second in the duration of testing. Finally, the board-to-board connector was successful in flashing the microcontroller unit (MCU) and the Bluetooth Low Energy module while the microUSB port facilitated successful uploading of Arduino code to the MCU.