2017Fall-IntelligentSecurityGuard

Quick link

• Webpage
• Motivation
• Requirements
• Implementation
• Installation & Usage
• Build Instructions
• Recognize Usage
• C Library in Python
• Result & Reference

Motivation

在這個房價高漲的時代，不是人人都能住大廈豪宅，就算負擔的起，每年的管理費也是十分驚人!因此，選擇無大廈式管理的居住環境已然是常態，而為了補足這些人居住安全的需求，有台自動機器巡邏車在家中顯得格外重要。我們期望設計出低成本、架構簡單的系統，取代一般警衛的工作，不僅省去人力成本，也保障使用者基本的人身與財產安全。

Requirements

Software

• python 3
• pexpect(python module)
• darknet-nnpack(already built)

Hardware

• Rpi3 model B
• Arduino uno
• Rpi自走車套件
• 行動電源(5V 2.1A)
• 9V電池
• 高亮白LED
• Rpi camera module
• 人體紅外線感應器navigation
• 單路relay
• 蜂鳴器

Implementation

See in Webpage

Installation & Usage

Upload replace/Arduino_car.ino to Arduino UNO.

Log in to Raspberry Pi using SSH.

git clone https://github.com/NTUEE-ESLab/2017Fall-IntelligentSecurityGuard.git
cd 2017Fall-IntelligentSecurityGuard
python main.py

For the first time, use Set Up Route to set up the route. Press GO! then the car will move along the route.

Build Instructions

If there is any problem, you can try to build the darknet yourself. But do check the directory path of dartnet-nnpack and that in recognize.py. Or the recognition might crash.

Log in to Raspberry Pi using SSH.
Install PeachPy and confu

sudo apt-get install python-pip
sudo pip install --upgrade git+https://github.com/Maratyszcza/PeachPy
sudo pip install --upgrade git+https://github.com/Maratyszcza/confu

Install Ninja

cd 2017Fall-IntelligentSecurityGuard
mkdir alter
cd alter
git clone https://github.com/ninja-build/ninja.git
cd ninja
git checkout release
./configure.py --bootstrap
export NINJA_PATH=$PWD

Install clang

sudo apt-get install clang

Install modified NNPACK

git clone https://github.com/shizukachan/NNPACK
cd NNPACK
confu setup

If you are using Rpi3, use the following commands. Replace NNPACK/src/init.c with replace/init.c

Otherwise, modify NNPACK/src/init.c to fit your CPU arch.

Also, add -fPIC to cflags and cxxflags in NNPACK/build.ninja.

$NINJA_PATH/ninja
bin/convolution-inference-smoketest
sudo cp -a lib/* /usr/lib/
sudo cp include/nnpack.h /usr/include/
sudo cp deps/pthreadpool/include/pthreadpool.h /usr/include/
cd ../

Install qasm2

sudo apt-get install flex
git clone https://github.com/Terminus-IMRC/qpu-assembler2
cd qpu-assembler2
make
sudo make install
cd ../

Install qbin2hex

git clone https://github.com/Terminus-IMRC/qpu-bin-to-hex
cd qpu-bin-to-hex
make
sudo make install
cd ../

Install qmkl

sudo apt-get install cmake
git clone https://github.com/Idein/qmkl.git
cd qmkl
cmake .
make
sudo make install
cd ../

Install darknet-nnpack

git clone https://github.com/shizukachan/darknet-nnpack.git
cd darknet-nnpack

Replace darknet-nnpack/Makefile with replace/Makefile then make it.

(Note that this Makefile is only for Rpi3. If you want to run on other platform, modify -mcpu=cortex-a53 part in the Makefile to fit the CPU arch you are using.)

Ref: shizukachan/darknet-nnpack

Recognize Usage

If you just want to test the human detection, use the following commands.

# initialize
from recognize import *
p = init()
# take a photo
shot()
# recognize(return true if detected)
recognize(p)
# move the prediction image
move()

The final image will be data/predictions.png.

C Library in Python

The origin version of darknet provides darknet.py for python user. Since now we build it with NNPACK, the script is no longer usable. After some studies, I make a new version. You can put replace/darknet.py under your darknet-nnpack directory and simply run it.

Result & Reference

See in Webpage