You are on page 1of 9
4. SUMMARY AND CONCLUSIONS In summary, the sonar unit proved to be extremely capable of fulfilling its role as a closed loop feedback detector. The main limitation of the system was the turtle which did not lend itself to having anything extra mounted on it. In the end the precariousness of the sonar mount eliminated any chance of having the turtle perform even basic maneuvers The most important part of the project turned out to be the development of the sonar-control ASSEMBLER program. Any fu- ture setup of the sonar could consist of a 6502 microproces- sor, a 6522 VIA with associated circuitry, the interface collection of transistors etc., and the power supply. This is opposed to the clocks, binary counters and such on the EDB (see appendix A). For one thing the units of the output are no longer limited to feet. Since a microprocessor chip is being used it can be made to do more than just give a range. Given this setup, there are many avenues of explora- tion. I would suggest that in any future work using the unit with robots, that a different robot be used. The robot should specifically be built with a platform for holding different devices. | The sonar unit could then be mounted like a radiotelescope on a stepper motor. Thus, the trans— ducer could be rotated instead of the entire robot. Lastly remote control would be preferable to having a twenty foot umbilical cord. If the unit was to be used strictly as a ee range finder, an acoustical horn could be mounted on the de- tector to focus the sound waves. There might even be a way to provide “stereoscopic vision". The trick would be to achieve this without using two so- nar units. There may be a way to do it with the two trans- ducers supplied in the kit. The sonar could indirectly con- nect to both transducers through an integrated circuit of some kind. The transducer used would be determined by the microcomputer. This would require that the transducers were both at the same angle with respect to some fixed position. The result would be much like the triangulation obstacle de- tection system mentioned in the introduction. It also might be helpful to decrease the lower range limit. The method listed in the instruction manual is to decrease the width of the ultrasonic pulse. This is not trivial. The major part of the work has to be in software. There is no reason why the robot should have to jerk along to its target. The controlling program must be made faster. It could simply be written in ASSEMBLER but multiplication and division subroutines would have to be written. The gain in speed is compensated for by a loss in sructure and readabil- ity. Actually though, these may be eliminated once it is realized that the computer can think in terms of time as easily as distance. The conversion to feet is done simply as a convenience to the user. A At measurement could be sent to the ASSEMBLER program instead of a distance. This would eliminate the extra conversion arithmetic. Once the sonar unit is put under software control, the range of things it can be used for increases inmensely. If a moving obstacle was moving slowly and/or very far away a collision avoidance system could be implemented. Tf things are set up in such a way that all that has to be done is recognize a change in height or width, a sonar unit could replace a television camera. In short, any application re- quiring that range be taken can be realized with the micro- computer controlled sonar unit. Appendix A EXPERIMENTS WITH THE SONAR UNIT Since it came already attached to the sonar unit, the ex- perimental demonstration board was used to provide the dis- tance readout for the experiments. For reference, a sche- matic of the ED8 is shown in figure A. The first experiment was to test the accuracy of the unit. This was done by mov- ing an object back and forth along a ruler that had the so- nar transducer at one end. ‘he results of this test showed that the unit could range objects with no discernible error. Another good feature was that the range to obstacle never changed once it was found (i.e., the number given did not change from cycle to cycle). the next experiment tested the resolution of obstacle detection. Thinner and thinner ob- jects were placed in front of the transducer until the dis- tance measurement showed that the object was no longer de- tected. ‘this limit was never reached. Apparently the unit can discern objects only a few millimeters wide. ‘The last test yielded the angle of detection. An object was moved left and right perpendicular to the wave direction. ‘The re- sult was an angle of detection of about +15 degrees. one other observation was that holding the transducer close to the table produced readings of .9 feet even when the trans~ ey ducer was pointing parallel to the table. This indicated e that the table was reflecting sound waves. This eliminated the idea of mounting the transducer close to the floor. 0 el . i ome | “eine? ia ae rs a | ° levee Event “gece ee | eer - ie = ie I Seen epee elaty (cope ieee To cece wat Lzmcne/oete [Scannng vind winged Be odes ° foes los Se ‘BED -t0- Seven DiSPAY Lt Sus Sega Appendix B SONAR ACTIVATION AND TIMING PROGRAM 51 FL DELAY LPL Lp2 ADTIME EDA #$01. STA $A001 LDA $502 BIT SA001 BEQ XLG JSR ADTIME LDA #$04 BIT $A001 BEQ FLG LDX $A008 EDA $A009 STA SOCFE TXA STA SOCFF sec sac #504 BCS ADDCON cue EDA #S01 ‘ADC SOCFE STA SOCFE LDA #826 Le ‘ADC SOCFF STA SOCFP LDA #$00 anc OCF STA SOCFE sec LDA #SFF SBC SOCFF STA SOCFF LDA 4SFP SBC SOCFE STA SOCFE JSR DELAY LDA #$00 STA SA001 JSR DELAY RTS LOX #804 LDA §SEF STA $A008 STA $A009 LDA $520 BIT $A00D BEQ LP2 DEX BNE LPL RTS. LDA §SFF STA SA008 STA $A009 RIS send VSW pulse wait for xLG activate timer wait for FLG get contents of timer Jow see text correct error in time get total elapsed time; store in memory for retrieval by BASIC program let VSW high for 200 ms and low for 200 ms Appendix C CLOSED LOOP CONTROL BASIC PROGRAM initialization 5 POKE 40962,255: POKE 40960,0: POKE 40963,1: POKE 40961,0 14 POKE 4,0: POKE 5,143 rate= 1780 microseconds per foot of range, high and low are where the assembler lan- guage stores the result of the sonar 18 780: LOW=3327 20 326: PLACE=256 50 "INPUT DISTANCE TO Go" 60 INPUT ZC 70 Gosus 3000 set original distance that other distances will be compared to 80 DEST=DIST 90 POKE 40960,6 turtle forward (see table one: 100 GosuB 3000 if the difference between the old range and the new range is greater than or equal to the distance needed to be traveled, stop the turtle and leave the program give the turtle some time to move 105 FOR J = 1 TO 100: NEXT J 110 IF (DEST-DIST)= > ZC THEN POKE 40960,0: GOTO 130 otherwise loop back 120 GoTo 100 130 sToP 3000 POKE 40960,0 stop motors 3002 DUMMY=USR(0) call to ASSEMBLER subroutine in ap- pendix A the delta t returned by the subroutine is turned into a dis- tance 3010 ‘TIME=PLACE*PEEK (HIGH) +PEEK (LOW 3015 POKE 40960,6 start motors 3020 DIST=TIME/RATE: RETURN ee is REFERENCE LIST Allen, Stephen A. and Rossetti, Tony. On Building a Light-seeking Robot Mechanism. Hanover, NH: BYTE Publications; Ines, 1978 Ciarcia, Steve. Home in on the Range!: An Ul trasonic Ranging System. Hanover, NH: BYTE Publications, Inc., November 1980. Hollis, Ralph. Newt: A Mobile, Cognitive Robot Hanover, NH: BYTE Publications, June 1977 Morris, Dave L. An Inexpensive Robot Guidance system. Cerritos, CA: Interface Age, April 1980 Polaroid Corportion. Ultrasonic Ranging system Description and Operation. Norwood MA: Polaroid Corp. Rockwell International. AIM 65 BASIC Language Reference Manual. North Brunswick, NJ: Rockwell, 1979 Rockwell International. AIM 65 Hardware Manual. North Brunswick, NJ: Rockwell, 1979 Rockwell International. Brunswick, NJ: Rockwell, AIM 65 User's Guide. North ‘979 Rockwell International. 6500 Programming Manual. North Bronewick, Nyt Rockwell thesenat torial 7 TO7S a9

You might also like