Programmable autonomous vehicles – Fundamentals, Part II

If we look inside ourselves and think about what makes us autonomous beings, we can conclude that those are two components: intelligence and senses. Everything we see, feel or hear, our intelligence processes in terms of making conclusions/decisions. If we want to give machines a point of view it’s clear that we have to equip them with senses – sensors – electro/mechanical devices that allow a certain level of autonomic behavior. The choice and practical realization of the sensors is the next step after choosing a motor, device driver and microcontroller, about which you had a chance to read in the first part about programmable autonomic vehicles.

Types and sensor selection

Like DC motors, sensors selection depends on the tasks that programmable autonomous vehicle completes. So we have: sensors for detecting objects in space (Infrared and Mechanical Sensors), sensors for measuring temperature/humidity of the room (Temperature and Humidity Sensors), sensors for DC motors (Encoders and Disk Sensors – Feedback Motor Control Systems) etc. If you want to bypass practical realization of the sensors, you have a wide selection of on-line stores with all the product details (http://www.robotshop.com/sensors.html). Besides the crucial task of orientation in space, the other task is regulating the speed of the driving DC motor, which will be discussed in this article.

Picture 1 - Mobile Robot Sensors

Mechanical sensors

Mechanical sensors detect objects by physical contact. They are used for realization of the orientation in space task and are simple for practical realization. The Micro Switches (hereinafter referred to as SW) used resemble insect tentacles. Picture 1c shows the electrical schematic of the connection between SWs and microcontroller (hereinafter referred to as MCU). At the moments of physical contact pull-up resistors R1 and R2 are connected to the ground, whereby the pins P3.2 and P3.3 respectively generate a logic zero. Conversely, pins are set to logical units (~+5V). Interrupt unit MCU is configured so that in time of decline from logic unit 1 to logic unit 0 (falling triggering edge), there is a triggering of the interrupt routine which guides the control unit to make a decision. The left and right SW make the front set of the mechanical sensors. A separate PCB is used in addition to the SWs also has tripolar screw connectors for coupling sensor signals and interrupt unit. PCB is installed directly in front of the front wheels (pictures 4 and 5). The back set of mechanical sensors is made of one SW for detecting objects while moving the programmable autonomous vehicle in reverse (picture 6).

Picture 4 - IR Sensor and front mechanical sensor

Infrared sensors

Infrared (hereinafter referred to as IR) sensor detects objects without physical contact. It is used for the task of getting around in space. Sensor is built of two components: photo transmitter and photo receiver which are installed side by side. In the picture 1.a1. electrical scheme is shown for the IR sensors based on IR LE diode (LD271 transmitter) and photo transistor (BPW77NB receiver). The transmitter constantly emits an IR wave (950 nm) which is reflected back to the receiver if there is an object in front of the receivers (picture 1.a2.), and vice versa ( picture 1.a3.)

The intensity of the collector current is inversely proportional to the distance between the photo transistors and reflective surface of an object. As closer the vehicle gets to reflecting surface, the collector current intensity rises, and vice versa. Therefore, the difference in potential between the collector and emitter is an analog signal that is unfit for direct connection with the MCU. And because of this the connection between IR sensor and MCU is realized by implementing a comparator – conversion of analog signal into digital – A/D convertor.

Comparator is made of operational amplifier (from hereinafter OpAmp) - LM324 IC circuit (picture 1d) . 2K potentiometer that is attached to the (-) input OpAmp defines reference voltage (for example +2.5V). Collector from the photo transistor is attached to the OpAmp’s (+) input. If the voltage on the (+) input is higher than the voltage on the (-) input, than the logic 1 is set on OpAmp’s output – conclusion, there isn’t an object in front of the vehicle. In the contrary, logic 0 is set on OpAmp’s output – conclusion, there is an object in front of the vehicle. In moments of transfer (on pin 1) from logic 1 to logic 0, the interrupt routine is triggered (EXT#0) which leads the controller unit to make a decision.

Picture 5 - IR Sensor and front mechanical sensor

The diagram in picture 1.a4. shows the conversion of the analog signal (1) into digital (2) compared to reference voltage (3). To make things easier, analog signal is displayed as a linear function of time. Note that by adjusting the reference voltage (3) we are also adjusting the voltage threshold i.e. sensitivity (distance between an object and the sensor) of the IR sensor. Additional configuring of sensitivity is achieved with rheostats Rb and Rc (implemented FT-63 trimmer potentiometers, 102 and 105 retrospectively) by regulating the intensity of collector power of the photo transistor. By increasing the resistance Rb, the sensitivity of the IR sensor is decreased to stat light noises.

The IR sensor is implemented on a separate PCB so it could use it for purposes of other projects. Finished PCB with strip copper lines and drilled holes has been used. LM324 IC circuit (14 pins – picture 1d) is not directly soldered to the PCB, I used the DIP-18 socket that was at my disposal. IR LE diode and photo transistors are soldered next to each other at the distance of 10 mm. The upper side of PCB is coated with black isolating tape before the soldering to reduce / eliminate the unwanted reflection of light (black surfaces don’t reflect light). Since IR light isn’t visible to the human eye, the implemented sensor can be tested with CCD chip i.e. with phone camera. Just point the camera to IR LE diode and you will see the ray of IR light (you can try it with the remote controller). The IR sensor is installed at the front side of the vehicle which completes the front set of sensors (pictures 4 and 5).

Picture 2

Synchronization of the front wheels and sensors

The front set of the sensor is physically attached to the structure of the front wheels instead for the supporting structure of a programmable autonomous vehicles, which is a difference in comparison to a mechanical sensor on the rear (picture 6). A configuration like that allows a simpler implementation of the managing software. Let's see for example the situation in Picture 2. Suppose that the programmable autonomous vehicle is currently moving forward and the IR sensor has detected the object 1 (see picture 2a). At that point, we stop the vehicle and rotate the front wheels to the left, thus rotating the front sensors during which the IR sensor detects the object 2 (see picture 2b). Since the object 2 is an obstacle for the further movement, we change rotation to the right, wherein the IR sensor detects the free space of a further movement (Picture 2.c). In the case of coupling of the front set of sensors with the support structure of the programmable autonomous vehicle (picture 2.d), only the front wheels are being rotated, whereby the IR sensor still detects the object 1! This is a complex situation for making a decision / conclusion, whether there is an object in front of a programmable autonomous vehicles? If you want to make your life easier, use the situations shown in Pictures 2a - 2c as guide, and forget about Picture 2d.

Picture 8 - Rohm 547 sensor

Feedback sensors for DC motors

Feedback sensor allows measuring of the rpm of the DC motor in purpose of regulating it’s speed. In picture 1.b1. there is the electrical schematics, based on Rohm 574 photo-interrupter sensor that makes a photo transmitter (LE diode – 800nm wavelength) and photo receiver (photo transistor). The working principle of Rohm 547 sensor is shown in pictures 1.b2. and 1.b3. If we put a bumper between the transmitter and the receiver, the ray of light pointed to the receiver is interrupted by the bumper, thus the photo receiver stops to lead (the cut-off area). Otherwise the photo transistor leads (saturation area). If the bumper is installed on the shaft of DC motor, the frequency speed defines the frequency that interrupts the ray of light. As well as the IR sensors, the difference in potential between the collector and the emitter of photo transistor is an analog signal which is converted to digital using the comparator. As the interrupt unit MCU is set to trigger on the decreasing edge of the signal, the collector of the photo transistor is connected to (-) input on the OpAmp. The inverted conversion of the analog signal (1) into digital (2) in relation to the reference voltage level (3) is shown in picture 1.b4. Decreasing edge of triggering (4) represents the moment in which the bumper intersects the ray of light of the LE diode of the Rohm 547 sensor.

Picture 6 - Rear mechanical sensor

Pulse-Width Modulation

The DC motor speed control is done by regulating the voltage of the Vave (average voltage supply to DC motor) by using PW modulation according to the following equation: Vave = ton / T * Vin. Where: ton= active time of Darlington’s transistors ULN 2803 circuit, T = period of PWM signal, ton/T = duty cycle (often expressed in %) and Vin = power source (12V). For example: if the ton = T, the duty cycle is 1 (100%), that leads to Vave = Vin, which contributes to the maximum speed of the DC motor. In case when ton < T, for example by half, duty cycle is 0.5 (50%), which implies that the Vave= 6V during which the engine speed decreases. The larger the duty cycle is, the speed of the motor rises and vice versa. Active time, the period, the duty cycle and the PWM frequency are determined by the software (MCU timers) according to Rohm 574 sensor’s collected signals (feedback). Before implementing the software pay attention to the ton/toff time specifications required by ULN 2803 IC circuits which must be fulfilled in order for the transistors to be able to work stable (picture 3). If you want to bypass software implementation of the PWM signal, you can implement it by using the LM324 circuit.

Picture 3

For sensor electronics of the DC motor a separate PCB is used which is equipped with: LM324 circuit, trimmer potentiometer (FT-63 series, 504) for defining the reference voltage of the OpAmp and pin-head connectors used for easy assembly/replacement of the components (picture 7). PCB is installed vertically on the supporting structure of the autonomous programmable vehicle and parallel with ULN 2803 device driver. The Rohm 574 sensor is installed on the bottom of the vehicle. On DC motor’s drive shaft a bumper is installed for interrupting the ray of light (picture 8). For stability, Rohm 574 sensor must be installed so that the bumper engages the whole ray of light. The hardware part of the vehicle is completed by practical implementation of the sensors. The final step is implementation of the managing software which guides us into the domain of state machines with final number of states and ‘infinite’ number of transitions between them, which will be discussed more in the third part.