Project B.I.N.S.

Senior Design Project:

CDR

Introductions, the Encore to an Encore

Professor Tessier

Advisor

James Doty

Sensor Subsystem

Belief Gratini

Server Subsystem

Marcus Mei

Power Subsystem

Frank Zhang

Firmware

Subsystem

• Overflow of trash is an environmental and health hazard for parks, cities, and

schools.

• Waste management resources are limited, so we can find ways to better optimize

them.

Problem Statement

Barrel Integrated Networking System

As a real time, automated, trash-sensing system

B.I.N.S. will:

• Monitor trash levels

• Provide notifications on fill level

• Be self sustaining

Learn more at: http://www.ecs.umass.edu/sdp/sdp21/team01/

Our Solution: Project B.I.N.S.

1. Capable of operating from 0°C to 50°C.

2. Capable of operating without sunlight for 1 week.

3. Placement of unit will not impede use of trash bin.

4. Ultrasonic sensor capable of measuring up to 28 in.

5. Sensor capable of making measurement within 10 seconds, accurate to 2 in.

6. Connect to a wifi router within 100 feet obstructed.

7. Weather resistant housing for sensor unit.

8. Should not exceed $100 per unit.

System Specifications

Block Diagram

1. How will the device exist in a trash bin without impeding use?

2. Details on WiFi timing and effects on power/battery life.

3. Decisions on machine learning and location services.

Steps towards addressing these (Discussed in full ahead):

1. Design of self-attaching, minimal obstruction housing for final PCB system.

2. Determination of timing based activities and battery-life estimate for integrated

prototype.

3. Reevaluation of benefits of machine learning and location services in this

project.

MDR Concerns

The Physical Structure

Estimated Energy Consumption

Integrated Prototype:

• System draws ~140-160 mA awake and

transmitting

• Our battery capacity is 500 mAh.

• Worst-Case Timings (Measured)

• Wake-up to AWS connect - 17 sec

• Ultrasonic measurement - ~8 sec

Worst-case assumption that the system wakes up at peak

current for 30 seconds, 4 time per hour

Why this can improve:

• Switching regulator to buck 8.4V batteries down

to about 4.2V before final LDO

• Based on datasheet efficiency values, results in

~40% reduction in current draw by most power

hungry component

• Migration from Arduino Nano to AVR MCU

could cut current consumption by upwards of

20mA

• Potential to utilize deep-sleep RTC memory

not transmit repeats

Battery-Life Estimate:

~4 Days

Our Integrated Prototype

Our CDR Deliverables

Power Subsystem

Sensor Subsystem

Server

Subsystem

Firmware

Subsystem

The Current Prototype - Hardware Overview

Level Shifting

Rx Comparator

WiFi Enabled

MCU

5V MCU

Tx Driver

Rx Amplifier

A breakdown of the firmware and sensing

subsystem hardware

Charge

Controller

3.3V

Regulator

Solar Protection

A breakdown of the power subsystem hardware

The Current Prototype

- Solar Performance

Notable Accomplishments since MDR

● Full charge cycle with solar panels

● The Constant Current phase and

Constant Voltage phase can be

verified by inspection

● Functioning of switching regulator

circuit to supply WiFi MCU

CC CV

The Current Prototype - Hardware Performance

Target Location (in)

3 8 13 18 23 28

# of Measurements Taken

514 966 631 116 354 474

Avg Time to Measure (s)

0.6053 0.3238 0.5144 2.6919 1.0121 0.747

Ratio of Erroneous Data (%)

0 0% 0% 24% 22% 16%

Notable Accomplishments since MDR

● Powering of sensor and WiFi MCU

with battery driven regulators

● Logic level shifting enables

communication between US sensor

and WiFi MCU

Notable Accomplishments since MDR

• WiFi MCU now has a function to call the

ultrasonic sensor

• Takes 10 measurements, outputs mode of

those measurements

• Currently occurs after connecting to WiFi,

AWS

The Current Prototype - Software

• Deep Sleep Mode w/ Periodic WakeUp (left off

prototype for demo purposes)

• Storing the measurement in RTC memory

• Comparison to EE_PROM

• Not using Machine Learning or Location Services

The Current Prototype - Software

Previous Prototype built

• Login in to use app

• Register users to use app

The Previous Prototype - Application

Notable Accomplishments since MDR

• Main features built (Core)

• Sensors

• Display Fill, Battery, Solar, ID

• About Us

• Information about BINS.

• Home

• Sensor Pairing and Start Up Instructions

The Current Prototype - Application

AWS - Backend Architecture

• Sensor -> IoT Core -> Mobile

• BINS Uploads to IOT Core

• IoT Core Topic is obtained by Lambda Feeder

• AWS Sync Stores Data in Dynamo DB

• AWS Sync pushes Data to Mobile App

The Current Prototype - Architecture

Demonstration in Progress:

Pin the “Demo Camera” and “Frank’s Shared Screen”

Logistics Moving Forward

Custom PCB - Description

What was our goal?

● To design a hub to contain all practical

hardware that seamlessly interfaces with

components required to be off board (i.e. solar

panels and batteries)

What do we have?

● This board has been fabricated and is in

hand

● 10 mil trace widths provide sufficient

conduction for all supplies and signals

● Packs all components into an organized 2

inch by 4 inch package

Fabricated custom circuit board for the B.I.N.S system (Full schematic can be found in the

appendix)

Custom PCB - Layout

ESP-Wroom-2

module for reference

Pictured is a model of our PCB from Altium

Designer, with designators to breakdown how

each stage of our block diagram fit on board

• Assembly and testing of custom PCB

• Finish migration from Arduino board to AVR MCU (delayed by unforeseen

circumstances)

• ICSP (via UART) of the WiFi MCU

• Print and construct system housing unit

• Design verification according to system specifications

• Implement testing plan and intuitively display findings

• Finalize mobile integration

• Complete implementation of aws client

Steps towards FPR

Projected Unit Cost

Per Unit Price

2 x 3.7V Lithium Ion Batteries 500 mAh $15.90

2 x 2.5W Solar Panels (w/ Connector) $12.99 ($2.18)

Dual Cell Charge Controller (BQ2057) $2.55

4 x Schottky Diode (MBRA210LT3G) $1.52

3 x Voltage Regulators $3.56

4 x Transistors (3 MOSFET, 1 BJT) $0.93

Thermistor NTCLE100E3103JB $0.69

RS-232 Tx Driver (MAX203E) $8.17

Rx Amplifier and Comparator $0.65

2 x Ultrasonic Transducers $13.46

ATtiny85 Microcontroller $1.63

ESP-Wroom-2 (WiFi MCU module) $2.60

Estimate of PCB and Res/Caps $3.50

Estimated Per Unit Cost:

$70.33

$190 remaining in budget (see appendix for

breakdown.

James Doty: Hardware Lead, Sensor Subsystem

Belief Gratini: Software Lead, Server Subsystem

Marcus Mei: Team Coordinator, Power Subsystem

Frank Zhang: Budget Lead, Firmware Subsystem

Project Management

Gantt Chart

Questions? Comments?

Concerns?

Appendix

Inclusions per CDR Rubric

Custom PCB Schematic

Hardware (PCB Equivalent if not same):

• 2 x 3.7V Lithium Ion Batteries 500 mAh

• 2 x ALLPOWERS 2.5W Solar Panels

• 4 x Schottky Diodes (MBRA210LT3G)

• BQ2057 Dual Cell Charge Controller

• LM7805 5V Regulator

• LM1086 3.3V Regulator (TLM760M33)

• LM2575T-ADJ Adj. Switching Regulator (TPS5602)

• IRFD9010 P-channel MOSFET (RQ5E5025)

• NTC 10k Thermistor

• MAX233 RS-232 Driver/Receiver (MAX203)

• LM2902N Quad OpAmp (LM324D)

• UT-1240K-TT-R Ultrasonic Transceiver

• UR-1240K-TT-R Ultrasonic Receiver

• LM311 Differential Comparator

• Arduino Nano (ATtiny85)

• ESP8266 Huzzah Module (ESP-WROOM-2)

List of Hardware and Software

Software:

● Arduino IDE

● Amazon Web Services (AWS)

● AWS Amplify

● AWS Dynamo

● AWS Synch

● Google Cloud Pub/Sub

● Agilent Data Logger

● Onshape

Project Expenditures