Difference between revisions of "F17: Tata Nano"

Parts List & Cost

DBC File Link

Link to the DBC file which defines the CAN communication of the system DBC link on GitLab

Sensor Controller

Design & Implementation

Sensor controller is responsible for Obstacle Detection. This project is designed to use two sensor components, LIDAR, which is the state of the art sensing components that are being used by the self-driving car industry to map the objects in the vicinity of the vehicle. The other component is more traditional and tested approach for object detection, Ultrasound sensors. Employing the capabilities of these sensors gives and very robust sensing system for the self-driving car. This also ensures to cover the flaws of each system.

Hardware Design

The Lidar Sensor communicated to the SJONE board using UART communication protocol for data collection. The Lidar Core and Lidar Motor required seperate power sources and so two different 5V sources were connnected to the Lidar Sensor. Lastly, the Lidar Sensor required a Motor Control Signal to control the RPM of the motor. This was simply set to the high state by connecting it to 3.3V from the SJOne Board. Setting this signal high represented setting a 100% duty cycle on the Lidar and would set the highest RPM possible on the Lidar.

Lidar Schematic

Hardware Interface

In this section, you can describe how your hardware communicates, such as which BUSes used. You can discuss your driver implementation here, such that the Software Design section is isolated to talk about high-level workings rather than the inner working of your project.

Lidar Sensor

Lidar Sensor is interfaced using a UART Protocol. The Rplidar System has several op codes used for communication and control of the system. The Rplidar Op Codes and C code implementation is shown below.

From the commands above the majority of their use was for troubleshooting, testing, and to determine how the data flow/processing on the SJOne would be handled. From the three types of scans that are listed above it was decided to use the standard scan. In the images below we can see the continuous response mode that the standard scan follows after a scan start request has been sent.The first packet sent after a start command is a response descriptor which confirms that the lidar has received the command and will begin sending 360 data this can be seen in the center picure. Lastly 360 lidar data is sent in 5 byte data chunks. The contents of the 5 bytes is shown in the leftmost image below.

As shown below the Lidar Sensor sends 5 bytes per data point over UART to the Sensor Controller. These 5 bytes include 1 byte for the quality of data, 2 bytes for the angle of reading, and 2 bytes for the distance of reading. The Sensor Board initiates the Lidar Sensor readings by sending a start command opcode. The Lidar Sensor responds by sending a start scan confirmation followed by continuous data readings.

Rplidar

Ultrasonic Sensor

LV‑MaxSonar‑EZ1 ultrasonic sensor by MaxBotix is used for a wide range object detection. Ultrasound sensors are configured as 2 ultrasound sensors in the front and one in the rear as the initial test configuration, this configuration helps to work in conjuncture with Lidar sensor mounted on the center of the car. The final project makes use of one ultrasound sensor that is dedicated for the detection of objects in the front of the car. LV‑MaxSonar‑EZ1 can detect objects from 0 inches to 254 inches, the object detected within 0-6 inches are provided with range information of 6 inches and the resolution is 1 inch. LV‑MaxSonar‑EZ1 provides three output formats pulse width output, analog output, and RS232 serial output. This project is using pulse width as the output from ultrasound sensors.

The following figure shows the pinout of the LV‑MaxSonar‑EZ1 ultrasonic sensor.

Sensor Pin Out

Software Design

Show your software design. For example, if you are designing an MP3 Player, show the tasks that you are using, and what they are doing at a high level. Do not show the details of the code. For example, do not show exact code, but you may show psuedocode and fragments of code. Keep in mind that you are showing DESIGN of your software, not the inner workings of it.

Implementation

This section includes implementation, but again, not the details, just the high level. For example, you can list the steps it takes to communicate over a sensor, or the steps needed to write a page of memory onto SPI Flash. You can include sub-sections for each of your component implementation.

Testing & Technical Challanges

Motor & I/O Controller

The Motor and I/O controller board is responsible for control of the motors and relay information between the LCD and the rest of the System.Hence, the board could be divided into 2 parts, the motor control, and the LCD control. The motor control logic controls the steering and the speed of the vehicle using a servo and DC motor respectively based on the Car Control CAN message from the master. Based on other can messages from the Sensor Board, the Geo Board and the Android Board, the Motor and I/O board processes and communicates the information to the LCD display

Design & Implementation

The Servo and DC motor are controlled via PWM and hence initial work required finding out the required duty cycle values for desired motor frequency. Higher frequency gives us a better resolution and response from the motor but for this project, a frequency of 8Hz was ideal enough to attain the required speed of vehicle and response time of the steering of the vehicle. To find out the PWM values, we connected the RC receiver of the Traxxas Slash 2WD vehicle to an oscilloscope and varied the remote controller for forward and reverse movement as well as right and left movement of the car. PWM signals were observed on the CRO and as the controller trigger was varied, the duty cycle of the PWM signal on the CRO also varied.

Speed control of the vehicle was carried out using a Traxxas speed sensor and a bunch of magnets. Applying the principles of a hall effect sensor, the magnets were attached to the inside of a wheel and the speed sensor was placed on the shaft of the back wheel. With every rotation, the magnets cut the field of the speed sensor giving a positive voltage to the SJone board. This positive voltage is accounted for and after neceassy calculations, we derive the speed of the vehicle.

To read data on the fly, LCD display by 4D systems was interfaced. The SJOne board communicates with the LCD display over UART using basic ASCII values that represent commands as well as information. The LCD graphics are preprogrammed into a MicroSD using the 4D systems workshop software and each graphical object consists of ASCII commands to control it.

Hardware Design

The Motor and I/O controller system consists of the following modules to perform various functions as mentioned in the description section of the table

Hardware Interface

Electronic Speed Controller (ESC)

The ESC is the interface between the DC motor and SJOne board. The ESC enables speed control, protects the rest of the system from any back EMF and allows configuration of the motors in various Modes (Training/Race/Sport). The ESC has 2 connectors the first one is a 2 wire connector, black and red which is connected to the LiPo battery that powers the motors and the second if a connector of 3 wires. 2 wires (black and red) supply a 7V DC power stepped down from the 11V lipo battery to the motors. This 7V power is used to control the servo motor using the power distribution board designed for this project. The 3rd wire (white) is a PWM input signal to the ESC from the SJOne controller that defines the speed of the motor.The ESC consists of a button to calibrate and turn on/off the ESC located on it.The ESC can be calibrated by following the steps mentioned on the Traxxas website

Traxxas ESC XL5

DC Motor

Traxxas DC Motor

The speed and direction of rotation of the motor (forward/backward) are controlled by the direction and amount of current that is supplied to the DC motor. In the figure below you can see the motor has 2 wires; one for positive(Red) and one for negative(Black). For the forward movement of the wheels the current flows from positive to negative, and for the reverse movement, the current flows from negative to positive. The speed is controlled by the amount of current it is fed from the ESC which is in turn controlled by the duty cycle of its PWM signal input.

Servo Motor

Traxxas Servo Motor

The direction the vehicle's front wheels turn is dependent on the servo motor in the vehicle. Based on various PWM signals, the servo steers the front wheels of the vehicle in the left and right direction. The servo has 3 wires of which one is for the PWM input signal whereas the other two are to power up the servo. We powered our servo motor using the 6v power supply from the battery elimination circuit present in the ESC so that a single switch to turn on and off both the servo and DC motor.

Speed Sensor

Traxxas Speed Sensor

To maintain the speed of the vehicle, a speed sensor from Traxxas was used. The assembly provided a single magnet and required mounting the sensor in the rear compartment. This setup had a major drawback of using just one magnet. One magnet did not provide enough resolution for the speed check algorithm at low speeds and small distances. Hence, we opted to mount the speed sensor on the motor shaft and attached 4 magnets on the wheel. The sensor works on the hall effect principle where it provides a current across its terminal when it is close enough to a magnet. These pulses are read by the SJOne board and fed to the speed control algorithm. the speed sensor has 3 wires, the white where are the output wire that provides the pulses to the SJone board and the other wires power the sensor.

Magnets

uLCD32-PTU

uLCD32-PTU by 4D systems has a 3.2" TFT LCD Display module. The module comes with a display resolution of 240x320 pixels. 4D Systems provides a programming cable based on UART for burning the LCD code to the module. The project is burnt to a uSD card which is used for display during booting of the LCD. It is recommended that we use the programming adapter provided by 4D systems as it has a special reset button that can be used to download the built project to the LCD display. Using other programming cables like CP210X or another FTDI chip did not help in downloading the project to the LCD display module. Once the LCD display was configured with different widgets and screens, the motor module was coded to display information in LCD through UART (There is no need of reset button connection here as the motor does not have to send any reset signal).

The following figure shows the programming cable and the pins used for uLCD32-PTU.

LCD connections

Software Design

Motor and IO Controller

The motor control logic controls the steering and speed of the vehicle by providing duty cycles defined by set macros to the ESC and servo motor. These duty cycles were derived by reverse engineering the motors using the RC controller.

The motor control algorithm and speed control algorithm execute in the 10Hz periodic scheduler to provide a quick response at 100ms. That is, the motor algorithm is executed and responds every 100ms. Whereas the LCD logic is controlled in the 1Hz task as a refresh rate of 1s is sufficient to display information for debugging. The Speed macros are:

#define VERYFAST    18.0  //Avoid Using. 
#define FAST        17.0
#define MEDIUM      16.0
#define SLOW        15.6
#define VERYSLOW    15
#define STOP        14
#define SLOWREVERSE 12.6
#define REVERSEFAST 12

Whereas steering wheel macros are:

#define EXTREMELEFT 11 //Avoid using
#define EXTREMERIGHT 19 //Avoid using 
#define HARDLEFT 12
#define LEFT 13
#define SOFTLEFT 14
#define CENTER 15
#define SOFTRIGHT 16
#define RIGHT 17
#define HARDRIGHT 18

DC Motor and Servo Motor

The steering and speed algorithm is controlled by input can message from the Master which is decoded into steering and speed commands. Using software abstraction and encapsulation, the DC motor and servo motor are both objects that can be controlled by its SetSpeed and SetDirection commands respectively.The decoded can message is then passed as variables to both the steering and speed object to obtain the desired motion.

The steering and speed commands, as mentioned in the previous section, are macros we defined that represent PWM in terms of duty cycle that controls the servo and DC motor. We have defined 9 different steering angles and 6 different speed values. The speed values, however, are only initial signals that are then increased or reduced to maintain speed on slopes. In the various steering macros, we are avoiding extreme turns as we noticed the gears got noisy at those values.

10hz Task Flow Diagram

RPM Sensor

The RPM sensor along with the magnets calculates the total revolutions of the wheel and compares it to the desired value. Each speed value has a corresponding RPM count for the wheel and the calculation is done using the formula :

  speed = (Circumference * RPM_Cut_Count * 3600) / (1000*Constant);

Where speed is in Km/Hr and RPM_Cut_Count is the number of times a magnet cuts the RPM sensor. Circumference is the circumference of the wheel at 36.5 cm. Constant is a value that helps adjust the value to work with the 3 magnet arrangement. In our case, the constant is of value 300. The 3600 / 1000 is a constant that changes the value from m/s to km/hr

LCD DISPLAY

uLCD-32PTU communicates with the SJ1 board over UART at a frequency of 1 Hz with a baud rate of 115200 bps. In order to reduce the amount of data transmitted over UART frequently, the code checked scans for which page is active and sends only the data of that page for display. Certain critical conditions such as bus resets are updated in the code frequently but are sent for display over UART only when the corresponding page is active. Data to be sent is preprocessed to HEX code before transmitting. The steps taken for interfacing the LCD display with the SJ1 board is shown below:

1hz Task Flow Diagram

Implementation

Motor IO

• Creating a project using Workshop 4 IDE and programming the LCD display

1. After finalizing the design of the LCD's layout, a genie project was created using Workshop 4.
2. The layout was split into different forms (pages) and appropriate buttons and gauges were added for display.
3. uLCD-32PTU was programmed with the help of programming cable provided by 4D systems.

• Programming SJ1 Board for LCD Display

1. As raw data often cannot be displayed on the LCD directly, the values to be displayed had to be converted to appropriate byte-sized values.
2. Communication with SJOne board was established at the baud rate of 115200 bps.
3. Commands for writing data and reading acknowledgment for various gauges were coded in the SJ1 board.

Main Page

Sensor Page

Distance Page

Angles Page

GPS Page

Testing & Technical Challanges

A major challenge that was faced while interfacing the LCD with the SJ1 board was that the SJ1 board was frequently getting rebooted while sending data for all metrics at once. To counteract this problem, the metrics were split into different forms(pages) for display and the data belonging to the active form alone was sent. Another challenge was that LCD does not support display of data that is more than 4 digits in length. To support display of data such as GPS coordinates, multiple 4 digit display objects had to be used with data manipulation before sending over UART.

Geographical Controller

Geographical Controller is one the most important controller in the autonomous car which help it to navigate to its destination. This controller consists of Global Positional System (GPS) and Compass (Magnetometer) modules. These modules continuously update the position and orientation of the car with respect to geographical north and sends the data to Master, Motor and Android controller boards. We are using Adafruit Ultimate GPS module.

Design & Implementation

The GPS module used in this project runs on UART communication protocol. Its default baud rate was 9600bps. We configured it to work on 57600bps to extract data at faster rate. The GPS module works on NMEA 0183 standards which defines the electrical and data specification for communication between GPS module and its controller. We are using Recommended minimum specific GPS/Transit data (GPRMC) command. It provide us with three importatnt data which are Fix, Latitude and Longitude, required for localization and navigation of the car. The update rate of these data from GPS module is configured at 5Hz.

Following are the parameters which are useful in developing GEO algorithm -

Bearing Angle - Bearing angle is an angle between the line, made by joining two points, with respect to Geographical north. Here, two point that are considered are - First is current location of the car and second point is destination or the next checkpoint to be reached. This angle is calculated using Haversine formula.

Heading Angle - Heading angle is directly given by Compass module. It is an angle made by the current pointing direction of the car with respect to Geographical North.

Deflection Angle - The difference of Bearing and Heading angle gives the Deflection angle. This angle is indication for amount of rotation the car should make to reach its destination point in a straight line.

Distance to Checkpoint - This is the distance in meters between current position of car and the next checkpoint to be reached.

Hardware Design

The GPS module we are using is Adafruit Ultimate GPS module which has following features -

-165 dBm sensitivity, 10 Hz updates, 66 channels 5V friendly design and only 20mA current draw Breadboard friendly + two mounting holes RTC battery-compatible Built-in datalogging PPS output on fix Internal patch antenna + u.FL connector for external active antenna Fix status LED

Hardware Interface

Software Design

Implementation

Testing & Technical Challanges

Communication Bridge Controller

This is a part of a project where different technologies meet. To make our project more understandable and easily accessible, we need a user interface. Where a user or customer of the product can interact with the product easily and without knowing the technical complexity of it. We have decided to make an android application which communicates with the car and can show its current location, speed, heading direction etc. Here, we have discussed the design and implementation.

Design & Implementation

The main purpose of this module is to exchange data by using a wireless communication protocol. We had options like WiFi (UART to WiFi converter) or Bluetooth (UART to bluetooth converter). Here we made decision based on actual requirement and usability of that communication protocol. In case of WiFi it has advantages like long range, high speed and robust communication without loosing data packets. It can also connect to multiple devices at same time. Now, the actual requirement is to show useful data to user, where use is sitting inside a car(idle situation). Which doesn't require long range and multiple receivers like WiFi provides. So, we moved ahead with bluetooth which works fine and delivers everything we wanted.

Hardware Design

HC-05 module was chosen as a better fit for this project. It is a serial to Bluetooth converter with a very compact hardware design. It supports Enhanced Data Rate Modulation with complete 2.4GHz radio transceiver and baseband. It's noticeable features pertaining to the project are low power operation (1.8 - 3.6 V) and programmable baud rate. Different settings of the module can be configured using AT commands.

Android Studio is used for creating an interactive Android application, Android Studio is available by Google for free, It's an intelligent software which helps programmer or developer with little or no knowledge of Java programming language to develop a good piece of code with high code readability and re-usability.

Bluetooth Module

Overview of hardware implementation

Hardware Interface

Connections with HC-05

Connections with SJOne

Software Design

Here, we will discuss the software design approach for communication bridge. As discussed in earlier topics, this communication is bridge is to send and receive data asynchronously to the main processes running over car end or mobile end. Let's make the idea clear what exactly we need to send and receive over bluetooth and when do we need to do that communication. As Preet asked all the teams to connect to car without pressing any button. It should automatically discover the car and be connected. When it comes to the data, we want to send start and stop commands and application expects important sensor data to display to user. Software design is divided into 2 parts,

• Android application
• Micro-controller end with HC-05.

Android application

Android application is developed using Android Studio free software from Google. The application structure is divided into 2 sections. First to use mobile bluetooth APIs to connect with the car and second is to show car's actual position on Google Maps. Let's talk about this in details.

• Homepage (Main activity):

In android application development every page on screen is called activity. Here homepage is named as main activity. In this activity we check for present bluetooth status to decide, whether application should start searching for car or not. If not turn on the bluetooth facility. Make sure that we have provided the application bluetooth permission which building the application. It starts searching for discoverable bluetooth devices around the mobile device. It particularly looks for MAC id of the found device. If it matches with the desired MAC id, it proceeds and tries to create the bluetooth socket. Meanwhile it shows on screen, if it found a new device or connecting to a device or connected to car. If user turns of the bluetooth facility manually from slide-down menu, it handles that situation gracefully and asks if user really want to turn it off or it should retry. After successful connection creation, it helps user to navigate to next activity to get command over car.

Flow chart for home page

Asking user to turn on bluetooth-1

Asking user to turn on bluetooth-2

searching for car

• Google Maps (Maps activity):

When user reaches this activity, it is assumed that connection was successfully established. Here, google maps API works in background to load the map from internet. To use their resource, Google wants us to have an activated key. Which has 2 versions, first is debug key and another is release key. As name suggests debug key is for testing and debugging purpose, where release key is for final version. After map is ready, as assuming that connection is still alive, application presents input buttons for user to take command over the car. Google APIs are very easy to use and understand. To give the checkpoints and destination to the car we used an API which gives location of the point we tap on screen. Thus, by tapping on screen we create a list of checkpoints which later be shared with the car with start command.

When user gives start command, he also has to specify a desired speed for car. Valid input for speed is from 0 to 10 miles per hour. Values other than this is fixed to the nearest end values of speed range (0 or 10). But there is an exception in speed. If user wants to test the car which we call free run, speed input for that is 55. 55 as a speed tells car to free run and there is no destination. Flow chart shown here explains the work.

Flow chart for Maps activity

Flow chart for receiving data

Micro-controller end

The controller handles the communication between the Android Application and the Car through the HC-05 Bluetooth Module.

All communication between the controller and the App takes place in a 1Hz periodic task. 1Hz task allows enough time for a string with a high number of checkpoints to get parsed completely which might not be the case with other higher frequency tasks.

The communication depends on the Start signal sent from the App to the controller. The controller also receives the desired speed and selected checkpoints along with the Start signal. On receiving the Start signal, the controller starts sending checkpoints based on request from the Master. It also receives GPS and Compass data and sends it to the App to display it. All communication stops when the App sends the Stop signal.

A flow diagram of the 1Hz task is shown below:

1 Hz Task Flow Diagram

Implementation

Testing & Technical Challenges

Android application

• Establishing connection: When application building was in initial stage, we were trying to connect any random devices to check the capability of our code. We were able to connect with some devices, but other were either taking too long or rejecting the bluetooth connection request. Later then we focused on only one device HC-05 which we were using on car. It is working well for that.

• Application compatible to all: I figured out that even when we have given list of permissions to application, phone sometime doesn't allow random application to use resources for security purpose. Like in our case, application was connecting to the car perfectly but in maps activity, map wasn't showing user's own location. Which tracked down to the point that we need to go in application settings and give permission for GPS, network and fine location. Then it worked well for multiple devices. There should be some extra steps in application from developer, which forces user to give permission if not given automatically.

• Extra code: Initially to get the list of previously paired devices, we had written a function. Now, in a scenario where car is not around, and we try to connect to it, this function was always returning true as it found the car's device in paired devices list and tried to connect it. As connecting to a device is a blocking event, application freezes for a moment if not written in thread, our function was always getting stuck there and successful connection ratio was 10% of all the attempts as 90% of the time it got device in paired device list and tried connecting to it. Then we always looked around by keeping mobile in discovery mode. If it finds a device with matching MAC address, then it tries to connect.

• Debug key for Google maps: Debug key is a permission from google to use their maps service for free but only for debugging purposes. If we share the apk with debug key, maps will show up blank on screen. While working on android code with the teammate, I figured out that debug key generated by my friend, wasn't working for me (blank map), but release key worked fine. I haven't got the solution for this yet but I believe it is because the machines we use. Google takes some machine information to keep track who is using their services. I may try generating new debug key from my machine and test.

Micro-controller end in freeRTOS

Master Controller

Design & Implementation

The master, also the heart and brain of this design, ensures that all the other controllers perform the expected operations. This is achieved by sending a 'heartbeat' signal to the other controllers periodically. The controllers are expected to suspend their operation in the absence of this heartbeat signal.

The master controller is the central decision maker. It's working is briefly depected by the following diagram:

Interaction between master and peripherals

The master takes decisions based on the data it receives from the sensors. A flow diagram explaining the decision-making process is shown below:

Heartbeats sent by the master

Control flow of the master

Hardware Design

Hardware Interface

Software Design

The navigation algorithm

Implementation

Testing & Technical Challanges

Testing & Technical Challenges

Describe the challenges of your project. What advise would you give yourself or someone else if your project can be started from scratch again? Make a smooth transition to testing section and described what it took to test your project.

Include sub-sections that list out a problem and solution, such as:

Conclusion

Conclude your project here. You can recap your testing and problems. You should address the "so what" part here to indicate what you ultimately learnt from this project. How has this project increased your knowledge?

Project Video

Upload a video of your project and post the link here.

Project Source Code

• Gitlab Code Link

References

Acknowledgement

Any acknowledgement that you may wish to provide can be included here.

References Used

1. uLCD 32 PTU datasheet http://www.4dsystems.com.au/productpages/uLCD-32PTU/downloads/uLCD-32PTU_datasheet_R_2_1.pdf
2. Workshop 4 user guide http://www.4dsystems.com.au/productpages/4D-Workshop-4-IDE/downloads/Workshop-4_userguide_R_2_1.pdf

Appendix