search

shaneallcroft / hide_and_seek

Launchpad 2017 Hide and seek robot project by Hunter Allen and Shane Allcroft
GNU General Public License v3.0
1 stars 1 forks source link

Is it possible to simulate a mobile sensor network within gazebo or using ROS in general? #3

Open shaneallcroft opened 6 years ago

shaneallcroft commented 6 years ago

This is the place to post any info if either of us find information on said subject so we can stay on the same page. Also, if it's not possible to simulate, realizing this early on is key to moving forward and exploring alternate options for searching algorithms. If it's not possible and we can't figure out a great way to find a seeking path, we can fall back on hard coding tourist points ourselves (this is kinda lame though, cause I'd like it to be as dynamic and self reliant as possible, so we'll save this as a last resort).

Additionally we should meet sometime sooner this week so I can explain my algorithm in further detail than what my cam scanner chicken scratch does B ).

allenh1 commented 6 years ago

@shaneallcroft What exactly do you mean by "mobile sensor network?" If you mean have sensors, definitely. A great place to ask questions about Gazebo is the Gazebo version of ROS Answers.

shaneallcroft commented 6 years ago

Just saw this, I'll check there, thank you!

shaneallcroft commented 6 years ago

Ah! Also! Sorry i didnt even answer your question before, by mobile sensor network I meant its definition as its used in this paper: http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.12.8990&rep=rep1&type=pdf.

shaneallcroft commented 6 years ago

I've read some things on the gazebster answers forum, and I think what I want to do should be possible, but I'll have to do some work on my own to get the proper forces defined for all the walls and all the different nodes that'll be moving. Also I think its important that I have something more powerful than my laptop because oh jeez will this thing die if we give it more than one bot to simulate. I'm going home this weekend and upcoming week for thanksgiving. I'll see if I can bring my desktop back with me for the heavy lifting. :+1:

allenh1 commented 6 years ago

From the paper:

A mobile sensor network is composed of a distributed collection of nodes, each of which has sensing, computation, communication, and locomotion capabilities.

That's precisely what ROS is (nodes always have those capabilities, but frequently don't use all of them).

In other words, when your launch file spins up each ROS node, you create the components of your mobile sensor network.

Also I think its important that I have something more powerful than my laptop because oh jeez will this thing die if we give it more than one bot to simulate.

Mutliple-robot networks are not, however, very easy to do with ROS. This is not quite necessary, though.

I'll have to do some work on my own to get the proper forces defined for all the walls and all the different nodes that'll be moving.

I'm not sure why you need to do anything with the walls here... That seems a tad unnecessary.