The bigger the building, the more likely it is to find a spot where the internet is not working as expected. It would be nice to have a map that shows how good of a connection we have everywhere in the building, or how the temperature changes between the different rooms.
At McMaster's Wireless System Reaseach Group (WiSeR), I have been working to develop a solution to this problem, and I named it WiserBot.
WiserBot is a rover that drives around the building, makes a floor plan, and at the same time will collect information about the building, such as light intensity, temperature, and wifi strength.
The rover is build on the Arlo Platform this is an excellent platform choice because the final rover is able to carry up to 30lb, and it can move up to 4m/s with fully charged batteries
The lower level navigation and PID is programmed in C, on the Arlo Propeller board.
The higher level functions such as the Simultaneous Localization And Mapping (SLAM), is programmed on the laptop attached to the bot.
The system is taking advantage of the ROS library. This makes the communication between the different modules very easy to configure, and easy to modify.
Robotics Operating System (ROS) offers the capability to easily take advantage of the different libraries the community creates, and to quickly integrate them with your own projects.
Although the rover is doing a pretty good job, there is a lot of room for improvement, and a lot of areas that can be explored to make it more efficient.
For example, the one of the next items I am planning to work, on, includes a smarter way of navigating the buildings, right now, it just drives around, in order to discover as much of the area as possible.
I would like to make it, so it takes into account which areas we have the most data about. It would make sense to take a longer path, if it can collect more data about those areas.
Another device I will add to it, either to work in parallel with the Lidar, or to replace it, would be a Google Tango.
The Google Tango is a device that we currently have in the lab, and it has a couple advantages over the Lidar, the most important one would be a difference in price. The lidar is in the $8000, price range, whereas the Tango is in the $500 range.
By using the Tango, I would like to take advantage of the camera on it, that way I would be able to create more meaningful maps, for example mapping where doors are, or map the posters and announcements in the building.
Here is a map of the WiserLab that has been created by the rover:
As it can be seen, it is a bit 'rough' as the lab is not just an empty building, but rather it has many different objects that can interfere, like chairs, desks, and other boxes. So this approach, might require some image manipulation to fill in the parts where items were impeding with the mapping process.
Here is a much cleaner map, done of the actual building hallways, there are less items in the hallways.
It is interesting to see the ways the Lidar reacts in a real world scenario, and what sort of quirks arise from real data.
For example, in the hallway map, in the lower left corner, there are some readings which seem to go though the wall. There are two doors there, those doors have a gap underneath them, and it leads to the rover seeing under the door.
This summer has been a great experience and I have learned a lot about robotics, and about probabilistic programming.
It is really a powerful way of representing a model, and I can see how it would be used in other situations as well.
Thought the next year, I will continue working on the rover, and I will continue implementing different algorithms and try to make it work as smoothly as possible.
Feel free to contact me with any questions you might have, if you are building your own rover :)
I would like to make it, so it takes into account which areas we have the most data about. It would make sense to take a longer path, if it can collect more data about those areas.
Another device I will add to it, either to work in parallel with the Lidar, or to replace it, would be a Google Tango.
The Google Tango is a device that we currently have in the lab, and it has a couple advantages over the Lidar, the most important one would be a difference in price. The lidar is in the $8000, price range, whereas the Tango is in the $500 range.
By using the Tango, I would like to take advantage of the camera on it, that way I would be able to create more meaningful maps, for example mapping where doors are, or map the posters and announcements in the building.
Here is a map of the WiserLab that has been created by the rover:
Here is a much cleaner map, done of the actual building hallways, there are less items in the hallways.It is interesting to see the ways the Lidar reacts in a real world scenario, and what sort of quirks arise from real data.
For example, in the hallway map, in the lower left corner, there are some readings which seem to go though the wall. There are two doors there, those doors have a gap underneath them, and it leads to the rover seeing under the door.
This summer has been a great experience and I have learned a lot about robotics, and about probabilistic programming.
It is really a powerful way of representing a model, and I can see how it would be used in other situations as well.
Thought the next year, I will continue working on the rover, and I will continue implementing different algorithms and try to make it work as smoothly as possible.
Feel free to contact me with any questions you might have, if you are building your own rover :)
No comments:
Post a Comment