SCANMANv1 - BEAM SONIC MOTOR CONTROL
07/07/01 - wilf rigter
 

INTRODUCTION

Here is another project from the wishing well: ScanMan1 is  a new BEAM circuit for sound control of a  two motor wheeled  mobile platform. This project  was inspired by, and I hope articulates the ideas of,  many BEAMers current and past. Sometimes it takes a simple question to trigger a new idea which expresses itself in a new design and article where some hidden BEAM treasures might come to light. My most important design criterion for any BEAM project is to make the hardware technically as simple as possible without sacrificing the complexity of useful behaviors. Taken to its logical conclusion : Do everything with nothing. ScanMan1 is one of those projects that comes close but I suspect that still more can be done with less. Finally, the name ScanMan was chosen because it aptly describes the basic principle of this project of using a scanned command  menu selection system. ScanMan1 is not related to a now obsolete mouse like handscanner with the same name by Logitech.

DESCRIPTION OF OPERATION

ScanMan1 uses two commonly available logic chips, the 74HC14 hex Schmitt trigger and the non inverting 74HC241  (or  preferred 74AC241) octal buffer. In addition 3 transistors, 2 small gear motors  and a handful of passive components are all that are needed to control the motion of a 2 wheel mobile platform from a distance of several meters using  whistle or hand clap commands. The operator selects one of four desired directions indicated by the 4 direction LEDs  located on the front, rear and sides of the platform. The LEDS are scanned sequentially at 1 Hz and when the desired direction is indicated, the command can be executed with a short whistle or hand clap. Three of the commands (left, right, reverse) will  time out after 10 seconds and the fourth command (forward) overrides the other commands to resume normal continuous forward motion.  In addition to the remote commands, three tactile sensors (left, right and reverse) provide autonomous avoidance or recovery from collisions with obstacles using the same time out period before resuming normal forward motion.

 The ScanMan1 design  consists of five subsystems:

1. PWM speed control for adjusted for 75%  forward speed and 25% reverse and turning speed.
2. Tactile sensor conditioning
3. Command scanner with indication LEDs
4. Remote control interface
5. Sound processor/detector
 

Figure 1 - ScanMan1- Beam Sonic  remote control
 

PWM CONTROL

The PWM oscillator is a non inverting bicore using two 74HC241 buffers which has output signals whose pulse widths are  adjusted by the ratio of two R/C time constants to generate short duration high level and longer low level output pulses. In the forward direction the speed is controlled by the low level pulses which sets the forward speed at 75% of maximum. The outputs of the PWM oscillator are in phase and are isolated  with a pair of 74HC241 buffers used as motor drivers. The PWM and associated buffers have the active high enable pin permanently connected to +5V

TACTILE SENSORS

The motors are driven on each terminal by 74HC241 buffers forming a CMOS h-bridge. The available drive current is limited and therefore restricts it to be used with efficient low current gear motors. The buffers on the sensor side of the motor have normally high level outputs set by the sensor pull up resistors. The sensors are simple switches for detecting an obstacle forward left and forward right of the platform. When closed by a momentary contact, they discharge one of two monocore delay circuits setting the output of the sensor motor driver to a low level for about 10 seconds. Turning occurs as the direction of one wheel is stopped or reversed  while the other wheel continues forward. This will cause the vehicle to take a sharp turn left or right.  The reverse sensor detects head on collisions and uses two diodes to discharge both left and right turning delay circuits. If both the left and right sensors are tripped the platform will also reverse to back out of corners or a narrow coridor. The two delay circuits are based on a simple monocore design which use capacitively coupled positive feedback to produce rapid and efficient switching at the end of the timeout period. Their principle of operation is also essential to implement the novel minimal remote control interface which will discussed later.

COMMAND SCANNER

The command scanner is a simple PNCfree micocore using four Schmitt inverters which produces sequential active low pulses at the four outputs. When a scanner output is low, the associated LED will ON indicating the currently selected direction. This provides a visual menu of four sequential  directions which an operator standing by selects and executes a command by issuing a sharp sound such as a whistle of hand clap when the desired direction LED is ON. Enhancements of this scanned command concept are in the works and will be added in a future revision of this article.

REMOTE CONTROL INTERFACE

Three outputs of the scanner are each connected with a diode and a 10K series resistor to the outputs!! of the monocore delay circuits. Normally the monocore output is not affected by this connection but when the monocore buffers are tri-stated by an 100ms active high command enable pulse from the sound processor, this connection acts as active low wired OR  logic to set a new turn or reverse direction With the buffer output at high impedance, the floating output level is forced low by the relatively low impedance of the 10K resistor and diode.  This low signal is capacitively coupled through the 10uF capacitor to the input of the monocore buffer. When the buffer is reenabled, this active low input and output state remains until the delay circuit times out 10 seconds later unless overridden by the FORWARD command. When fourth scanner output is active low , its LED current supplies base current for an PNP transistor which inverts the scanner FORWARD output. The PNP collector is connected via a 10K resistor and two diodes and when the active high command enable pulse is received, the delay circuit outputs are forced high to restore forward motion. The tactile sensors take priority over the remote forward command and will unconditionally force turning or reversing direction while the contacts are closed. This scheme is presented here as a novel method to wire OR remote control signal to a tactile sensor conditioning circuit without adding a separate interface chip. Some of the  passive components can be replaced  if a second 74HC241 tristate buffer is inserted between the scanner outputs and the sensor delay circuits and in addition the 4 spare buffers can be placed in parallel with the motor drivers to increase available drive current.

SOUND PROCESSOR

The sound processor is a slightly modified front end of the BeamSonic circuit and uses a simple pulse output to enable direction commands at the end of a sound pulse with a sound level and frequency is sufficiently high to trigger the Nv and Nu filters. The active low output is indicated with a LED for the duration of the sound pulse at the end of which the circuit generates an active high tristate pulse. Two preamplifiers provide gain and high pass filtering of the electret microphone output. The mic should be sensitive enough to pulse the output  LED ON with a short whistle or hand clap from 3 meters away.  It needs to be carefully mounted e.g. surrounded by  foam plastic,  to avoid acoustic vibration from the platform coupling into the mic and causing  false triggers. Mind you  the occasional random triggers could produce some interesting and unpredictable behavior. Similarly a series 1K resistor and 10uF filter capacitor to ground may be needed for the preamp power supply if interference occurs from power supply noise. Again some amount of noise feedback may allow heavily loaded and struggling motors to initiate a  random escape strategy which would be better than being stuck on a steep incline going nowhere.

SUMMARY

The foregoing provides a detailed overview of the description of operation of the ScanMan1 circuit. I am hopeful that this will stimulate new ideas about practical BEAM control systems. I will continue to concentrate on improving the design and would be happy to get feedback from others who build and experiment with this ScanMan1 design. The several planned improvements for the near future will include a simple circuit for  photo tropic behavior, a photo phobia command and novel alternatives to the ScanMan1 command GUI.

enjoy

wilf