TexasMaps

From Texas Instruments Wiki
Jump to: navigation, search
TexasMaps Logo.png

Texas Maps is an indoor mapping solution, designed to be the next step of navigation.
Whether you get lost in ikea or struggling to find an office, Texas Maps takes you exactly where you want to go.

Authors

  • Tanfer Alan
  • Max Groening

Features

Texas Maps device is designed to measure distance, locate users*, transmit information to the mapping software and generate sketches.

  • Absolute distance measurement (accurate to 1 cm)
  • Data transfer via serial PC Interface

Texas Maps device includes 2 ultrasonic sensors (HC-SR04), a magnetometer (HMC5883L) and TI FRAM experimenters board (MSP-EXP430FR5739). Designed to exploit the accelerometer for user positioning* and measurement correction.

TexasMaps indoor sim.png


Instructions

  1. Connect the device to your PC via USB.
  2. Open a serial terminal (9600 baud, 8N1) **
  3. That's It! Now you should be able to see the unsigned integer samples taken
  4. Save the received files as mapLog.txt in to sample subdirectory of mapper3.tcl.
  5. Run mapper3.tcl
  6. Save the image of your model


Note: Weight, noise threshold, and smoother value can be easily adjusted to need.

TexasMaps runchart.jpg

Device

TexasMaps device 0.1b
Inside view of TexasMaps device 0.1b

A portable indoor mapping device that measures the distances to the sides via ultrasonic sensors, robust to user movements.
It is designed as a hand-held device with possible future wireless improvement.


Device works on the same principle as a sonar.

It bursts pulses to both sides sequentially, calculates the distance from the time delay between the pulse and echo.
After sending the pulse, it captures the echo with using capture mode of Timer B0.

It finds the time difference between sending the pulse and receiving the echo by simply multiplying this result with the speed of sound.

This gives 2 times the distance traveled. (1 for reaching the walls and 1 for reflection.)


Device transmits the measurement samples in centimeters through to serial port.


Using the accelerometer for user positioning is conceptualized in the following way:

  • Velocity can be found by adding up the accelerometer values (discrete summation Σ),
  • User position change can be found by multiplying the velocity with ultrasonic sample time.

This concept calculates the position changes by users movement such as walking forward-backward or unintended moves of mapping device during mapping. It enables to calculate forward walking distance per sample and to tolerate the left-right swings.


Using the magnetometer - the digital compass - is conceptualized to find users direction.
This is helpful because user can direct the device and sample anywhere without further sessions.


Ultrasonic connection
Trigger Echo
Sensor left P1.7 P1.4
Sensor right P1.6 P1.5

Processing

mapper3.tcl is what turns the distance samples collected by the sensors, into meaningful data below.

Incoming data is parsed according to the sensors they originated.
Noise reduction algorithms and fixes for sensor bugs are implemented.
Spiking sensor values are dropped with a software low pass filter routine.

It is highly customizable, user has an option to change the number of averaged samples. This gives a smoother (variables get smaller covariances) but less responsive model.
Wall thickness (weight) is adjustable to make the rare samples more recognizable.

Model of a small meeting room
Model of a coffee corner

These samples are actual outputs of mapper3.tcl script. From the figures above it can be seen that sensors do not collect any data before getting into the room. At this state, user is in the corridor out of measuring range, therefore there is no feedback (reflections) from any surroundings.

From the figure on the left the door can be detected as the user passes from the corridor to the room.

Model of a corridor

In the figure above, the start and end points are empty. They correspond to the stairwells.
On the corridor, open rooms cause minor reflections from the doors (little spikes), while passing by distance increases to corridor and room's combined length and gets out of the measurable range with current sensors.

Conclusion

Here is a combined model of the sampled areas.

Result of three samples combined and overlayed onto satellite image with a photo editor


Look at concrete labyrinths from the top. Texas Maps


*User positioning does not yet work properly
**RealTerm is recommended as the terminal program since the popular PuTTY or HyperTerminal does only support ASCII characters

Development code: File:TexasMaps0.1b.zip

Mapper script: File:TexasMaps Mapper.zip