Wednesday 14 September 2011

New Pics and Movies

An up to date picture of Zeus - Wires and LEDs worked in and sorted out :)

Heres a Video of Zeus walking around, at night, with his LEDs going.



 And one showing off his ability to rotate!

Friday 9 September 2011

LED systems

This robot will have two seperate LED systems, one will control his head LEDs, and the other his leg LEDs.
Ive taken some WIP pics of the circuits for them at the moment, so show off my (sub-par) soldering and heat shrinking skills.



Wednesday 7 September 2011

Speed Tests

My overall plan for this robot is to max out speed and agility, while having a higher than average firepower. (This is important due to weight considerations)

First part of this plan, is of course, to see just how fast he is at the moment:

Tests on 4 variable speeds (defaults):

Default walking modes are set by 3 variables - Body Height (range is -25 (body up) to 25 (body down)) - Leg Up Shift (range is 20 (close to floor) to 70 (about 2.5 inches from floor)) - Gait Speed (range is 3 (fast) to 20 (very slow))

L1: BH: 5 LUS: 30 GS: 3 --- 8m/60s - 0.13m/s (468m/hr)  (tile floor setting) (checked twice)
L2: BH: 0 LUS: 35 GS: 4 --- 8m/70s - 0.11m/s (396m/hr)  (default setting)
R1: BH: -25 LUS: 70 GS: 10 --- 8m/240s - 0.03m/s (108m/hr)  (tall grass setting)
R2: BH: 25 LUS: 35 GS: 4 --- 8m/87s - 0.09m/s (324m/hr) (low body setting)

Ive decided to use m/hr instead of km/hr or any other speed, mostly because it gives me a bigger number, and it makes it easier to compare between different speed settings.

So the fastest default setting was a slight body down, close to floor leg shift, and a fast gait speed. This was ~18% faster than the second fastest setting.

Next thing im going to do, is keep the fastest setting from the default settings, and then change the other settings manually, to try and maximise walking speed. (and hopefully not burn out any servos, or make the robot fall over himself)


L1: BH: 5 LUS: 30 GS: 3 --- 8m/57s - 0.14m/s (504m/hr) (tile floor setting - L1 from before)

L2: BH: 0 LUS: 45 GS: 3 --- 8m/53s - 0.15m/s (540m/hr) (average BH, average LUS, max speed) (checked twice)

R1: BH: 25 LUS: 45 GS: 3 --- 8m/68s - 0.12m/s (432m/hr) (low BH, average LUS, max speed)

R2: BH: -25 LUS: 45 GS: 3 --- 8m/60s - 0.13m/s (468m/hr) (high BH, average LUS, max speed)

As you can see in the results (I checked the fastest result twice) there is some inconsistency, since the L1 setting didnt change, yet was 3 seconds faster in this test. I was going to test again, but since I will be doing so anyway on the next test, I didnt. It appears an average body height is better than a low or high body height, at least in terms of speed.

The fastest setting here was average BH, average LUS and max speed. 540m/hr compared with the highest from the previous test, 468m/hr (ignoring the inconsistent result). This shows an improvement of  ~15% speed so far.


L1: BH: 5 LUS: 30 GS: 3 --- 8m/60s - 0.13m/s (468m/hr) (tile floor setting - L1 from first and second)


L2: BH: 0 LUS: 45 GS: 3 --- 8m/53s - 0.15m/s (540m/hr) (average BH, average LUS, max speed - L2 from second)


R1: BH: 0 LUS: 20 GS: 3 --- 8m/56s - 0.14m/s (504m/hr) (average BH, low LUS, max speed)


R2: BH: 0 LUS: 70 GS: 3 --- 8m/90s - 0.09m/s (324m/hr) (average BH, high LUS, max speed)

Well - these results are interesting. I was expecting with a low LUS that the robot would move faster, as it was spending less time lifting each leg. That doesnt appear to be the case. In fact, the robot began tripping over its feet as it was walking, which was probably slowing it down. High LUS definitely made him a lot slower, as he was lifting his legs very high. Interesting to note, however - he was 3 times faster with maximum leg lifting than the default leg lifting setting, because of the change in gait speed and body height.

Also, these results confirmed the inconsistency of the L1 results, I blame human error! :) I believe there are further grounds for testing however, as there may be a "sweet spot" for LUS, between 20 and 45. Anyway, I will leave that for another day, as I have found my fastest setting so far - average BH, average LUS, max speed - at a speed of 540m/hr. If I can decrease his time taken by even 2 seconds, that will boost his speed to 576m/hr, so thats my next goal in speed.

Other possible ways to increase speed:

*Bigger Servos may increase speed and torque
*Helium filled plastic tanks to decrease weight
*More efficient program code