DPRG
DPRG List  



[DPRG] Ideas for a Rover Class

Subject: [DPRG] Ideas for a Rover Class
From: Rud Merriam rudmerriam at gmail.com
Date: Tue Aug 5 14:58:04 CDT 2014

Doug,

I understand your points. The focus of the quadcopter class was to show 
that you can build the unit. It was to expand the number of people 
working with them. Basically, "You can build it and this is how." Adding 
more advanced features comes later. As you point out with rovers the 
diversity increases rapidly.

My goal is a rover that would be capable of RoboColumbus but only with 
addition of other components, e.g. vision or GPS. Going beyond that 
contest might require a totally new chassis but if someone becomes that 
interested I think they would be amenable to that extra cost. The 
original can always serve as a test bed for senors or software.


- 73 -
*Rud Merriam K5RUD
* /Mystic Lake Software <http://mysticlakesoftware.com/>
/

On 08/05/2014 12:07 PM, Paradug wrote:
> Rud,
>      This idea has a lot of merit. I do have some reservations about a 
> 2 day event. Also, currently we have people pursuing varied paths to 
> building an autonomous rover. I would like to keep this diversity of 
> approaches for at least through our November contest. A review of the 
> results of that contest would provide direction for the club’s next 
> step in developing a solid core of  skills.
>      On the other hand, an 6-8 hour event (maybe at RBNO or on a 
> weekend), that focused on skills that would be common to all 
> approaches might be a great kick in the pants for everyone. It might 
> also better fit in the busy schedules of members. It would be best to 
> poll the members and see if they are interested and what subjects they 
> would want included.
>      I am not too interested in a pushing a specific build for a 
> rover. I have seen members pursuing converted RC rovers, Dagu tank, 
> 6WD, and 4WD chassis , home built chassis, and David’s JBot.  People 
> are also investigating a multiple of sensor suites and the use of 
> ROS.  I think the different approaches are great.
>      Our immediate goal is for the rover to be able to successfully 
> complete RoboColumbus.  This means that the rover should be able to 
> move in a given direction for a specific distance, locate a target at 
> the end of the run and stop when it reaches the target. There are many 
> ways to achieve these four goals.
>      In the future, we want to our rover robots is to be able to 
> successfully run the Sparkfun Autonomous Vehicle competition, SRS 
> RoboMagellen, and successfully repeat David Anderson’s hat trick. This 
> would require the addition of good object avoidance, multiple waypoint 
> navigation, greater speed, and the ability to travel more difficult 
> terrain. If the rover is meant to evolve (i.e., not a RoboColumbus 
> specific robot), these things should be considered in the beginning of 
> the robot build.  However you must crawl before you walk. The time, 
> cost and skill restraints, that our members are under may require them 
> building a basic rover that can only meet the objectives of 
> RoboColumbus. A good intermediate approach is to build robust modules 
> with solid capabilities that can be transported to other robots.
>      Developing robust modules might be the approach to take at such 
> an event. As with your experience at the Houston TXRX makerspace, 
> people should leave the meet-up with a module that works. This means 
> that the hardware and software of the module is built to the level 
> that it can be used. Here is a quick and incomplete list of ideas:
> Sensors:
>     1. multiple ultrasonic sensor array capable of targeting an object 
> (similar to David’s JBot ultrasonic array).
>     2. Data merging GPS data
>     3. Targeting with Pixey
>     4. Neato LiDAR with ROS
>     5. Robust bumper system
> Communication:
>     1. A RC based deadman switch
>     2. Method to log run data
>     3. How to add FPV (real time or stored video)
> Navigation:
>     1. Build encoder and use with Arduino
>     2. Pololu IMU using DCM
>     3. APM2.6 autopilot
>     4. GPS S/W routines
> Other:
>     1. Behavior-based robotics with Arduino
>     2. Practical ROS
>     3. Blob detection and targeting with OpenCV or SimpleCV using Rp 
> or BBB boards.
> Again, the idea would be at the end of the session the participants 
> would have a working module that they could use on their robot. Parts 
> would have to be ordered and received before the event, there would 
> need to be tools to assemble the hardware and software that could be 
> used to drive the module. There would need to be a “teacher” who had 
> some knowledge about the goal and could get everyone started. Since 
> interest in a specific module would vary due the the individual’s 
> resources ($), strategy for competing, or just plain desire. The event 
> should maybe pick 2 modules to pursue, an easy one and a more 
> challenging one.
> Regards,
> Doug P.
> *From:* Rud Merriam <mailto:rudmerriam at gmail.com>
> *Sent:* Monday, August 04, 2014 2:52 PM
> *To:* Dallas Robotics <mailto:dprglist at dprg.org>
> *Subject:* [DPRG] Ideas for a Rover Class
> This last weekend I built a quadcopter at the Houston TXRX makerspace. 
> It was a 2 day activity, a lot of fun, and very intense. My brain was 
> fried by the end of the weekend. But it flies!
>
> I got thinking about doing a similar class for a an outdoor ground 
> rover. In line with some other comments recently about wanting to see 
> autonomous rather than RC robots the class project would be 
> autonomous, possibly with RC override.
>
> I am looking for suggestions for you in Dallas on how to proceed. If 
> we get the details worked out someone up there might like to give the 
> class at the DMS.
>
> A point I find surprising is how much more it costs to get an outdoor 
> rover chassis vs a quadcopter. The parts for the quad were around $150 
> but that did not include the battery or RC TX/RX. It did include a 
> micro-controller with accelerometer, gyro, and barometer. Other parts 
> include were the arms, plates, motors and props.
>
> I have some thoughts but let me gather any comments from y'all, first. 
> One important point is how much capability do you feel the rover would 
> need to be interesting? Keep in mind this is not the end-all rover to 
> win RoboColumbus but something to introduce people and show them they 
> can build and operate it.
>
>
>
> -- 
>
>
> - 73 -
> *Rud Merriam K5RUD
> */Mystic Lake Software <http://mysticlakesoftware.com/>
> /
>
> ------------------------------------------------------------------------
> _______________________________________________
> DPRGlist mailing list
> DPRGlist at dprg.org
> http://list.dprg.org/mailman/listinfo/dprglist

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://list.dprg.org/pipermail/dprglist/attachments/20140805/bc3e6e8b/attachment.html 

More information about the DPRG mailing list