dji matrice 100 usage with GPS plugged out - ros

I am using DJI Matrice 100 for indoor usage. I intend to use dji_sdk ROS library (http://wiki.ros.org/dji_sdk) to control the drone. The idea is to do path tracking. However, I do not want to use the GPS at all. General question: How do I get this done?
Specific questions:
How can I get the system to start with the GPS turned-off/unplugged? Someone here (Remove GPS on matrice 100) suggested that the latest update allows the system to take-off if the magnetometer is there but I do not find this to be the case.
I have a few options of controlling the drone by publishing on specific rostopics listed on the dji_sdk ROS page above. While I have successfully simulated this in the dji simulation environment, doing this in reality requires the direction of motion in the ENU direction to be known. Thus it requires the GPS to be ever-present. I do not know if there is a way to provide these commands in the the body-frame of the drone.
Please feel free to ask for any details you need!
Edit: For the 2nd part of the question, I was hoping someone would refer to dji_sdk ROS library (http://wiki.ros.org/dji_sdk) and tell me that I can use the rostopic /dji_sdk/flight_control_setpoint_generic to control the drone in the body-frame of the drone (as somehow referred to at the end of the page). If this is indeed true and using the flag you can achieve this, does someone know what should the value be for doing position control/velocity control in the body-fixed frame?

Related

Requirements to access IP Camera using OpenCV

I'm new to IP cameras and I know there are quite a lot of topics about this in the forum already, but I can't find a concrete answer for my needs.
I want to access an IP camera using OpenCV in Python from a Windows PC. As I don't have a camera yet, I need to buy one and I can't figure out, what requirements this camera needs to have.
For example, there are quite cheap IP cameras (e.g. Xi****) which say they come with an Android or iOS app and are only accessible via those.
I thought you can access any IP cam via OpenCV, but now I'm not sure anymore... can anyone give me an overview, what specs an IP cam needs, to be accessed via OpenCV on Windows? I don't want to buy a camera and later realize, that I can't access the video stream.
I'm really sorry, if this has already been asked, but I can't find a satisfying answer to this question and Google doesn't seem to be very helpful...
Thanks in advance.
check for IP cam that can transmit RTSP opencv know how to work with this type of stream.

Is this concept possible with iOS navigation tools

Lets say we have a starting point, (x,y). By using iOS navigation can we tell how far from that starting point we moved to another location (a,b). So if i walked 20 feet in a certain direction after starting would it be able to tell me how far I've moved and in which direction?
If this technology exists can I get info on where to start learning about it?
This also needs to be done without GPS, sorry.
As rmaddy mentioned With core location class and incorporating GPS in a project you can obtain a distance traveled by the person who is walking. I found a great step by step tutorial for you which has a sample project you can build and take a look at. Here is the link..http://www.perspecdev.com/blog/2012/02/22/using-corelocation-on-ios-to-track-a-users-distance-and-speed/
Also here is the link to core location class reference for further study....https://developer.apple.com/library/iOS/documentation/CoreLocation/Reference/CLLocation_Class/CLLocation/CLLocation.html
No, you can't determine location changes accurately without GPS, even with GPS it is difficult to accuracy measure position change as small as 20 feet (GPS 5m accuracy means a +/-15 foot error)
In theory you might be able to write software to create an Internal Navigation System using the built in accelerometers, gyros, and magnetometers, but in practice they are too noisy and have too much error for this kind of use (see this question). A better rocket scientist than me might be able to make it work but it was also need to use the GPS to keep it from drifting. The M7 chip on the 5S might make this feasible.

Face Tracking and Virtual Reality

I'm searching for a face tracking system to use in an augmented reality project. I'm trying to find an open source and multi-platform application for it. The goal is to return the direction where the face is looking to interact with the virtual environment, (something like this video).
I've downloaded the sources of the above Johnny Lee's application and tried to use Free Track too, making my own headset (some kind of monster, hehe). But it's not good to be limited to infrared points in your head.
These days I've download FaceTrackNoIR, but when I launch the program I get "No DLL was found in the Waterfall procedure." that I'm actually trying to solve.
Anyone knows a good application, library, code, lecture, anything that could help me to find a good path for this?
Thank you all!
I'll try to post results someday :-)
I would take a look at OpenCV. It is a general purpose machine-learning and computer vision C++ library. One of the examples in the download is a real-time face tracker that connects to a video camera connected to your computer and draws squares around any faces in the camera view.

Is there a virtual/dummy IMAQ camera for LabVIEW?

I'm writing LabVIEW software that grabs images from an IMAQ compatible GigE camera.
The problem: This is a collaborative project, so I only have intermittent access to the actual camera.I'd like to be able to keep developing this software even when the camera isn't present.
Is there a simple/fast way to create a virtual or dummy IMAQ camera in software? Ideally I'd like the dummy camera grab frames from an AVI or a stack of JPEG's. Something like this must exist, I just can't find it on Google.
I'm looking for something that won't take very long (e.g.< 2 hours effort) and that is abstracted away through the standard LabVIEW IMAQ interface, so that my software won't know or care whether its dealing with a dummy camera or an actual camera.
You can try this method using LabVIEW classes:
Hardware Emulation Using LabVIEW Classes
If you have the IMAQdx driver, you might consider just buying a cheap USB webcam for $10.
Use the IMAQdx driver (assuming you have it), and then insert the Vision Acquisition Express VI, and you can choose AVIs or even pics as a source.
Something like this: GigESim is a camera emulation software. Unfortunately it is proprietary and too expensive (>$500) for my own needs, but perhaps others will find this link useful.
Anyone know of a viable Open Source alternative?
There's an IP Camera emulator project that emulates IP camera with python. I haven't used it myself so i don't know if it can be used by IMAQ.
Let us know if it's good for you.
I know this question is really old, but hopefully this answer helps someone out.
IMAQdx also works with Windows DirectShow devices. While normally these are actual physical capture devices (think USB Webcams), there is no necessity that they have to be.
There are a few different pre-made options available on the web. I found using Open Broadcaster Studio and this Virtual Cam plugin to be easy enough. Basically:
Download and install both.
Load your media sources in the sources list.
Enable the VirtualCam stream (Tools > VirtualCam). Press Start.

Emulate GPS or a serial device

Is it possible to get location data out of Google Gears, Google Gelocation API or any other web location API (such as Fire Eagle) in such a format that it appears to other software as a GPS device?
It occured to me reading these answers to my question regarding WiFi location finding, on Super User, that if I could emulate a GPS unit, many of these web services could act as a 'poor-mans' GPS to otherwise less useful software that requires it.
Is GPSD an option?
Preferably OSX & Python, but I would be interested in any implementation.
There is a very similar thread on a Python mailinglist that mentions Windows virtual COM ports and discusses Unix's pseudo-tty capabilities. If the app(s) you want to use let you type in a specific tty device file, this may be the easiest route. (Short of asking the authors to provide a plugin API for what you're trying to do, or buying yourself a $20 bluetooth GPS mouse.)
Are you using OS X?
There is a project macosxvirtualserialport on Google code that provides a graphical wrapper around some of the features of a utility called socat. I'd recommend taking a look at socat if you see potential in the pseudo-tty route. I believe you could use socat to link a pipe from a Python program to a pseudo-tty.
Most native Mac apps will be querying IOServiceMatching for a device with kIOSerialBSDRS232Type, and I doubt that a pseudo-tty will show up as an IOKit service.
In this case, unless you can find a project that has already implemented such a thing, you will need to implement a driver as described in this How to create virtual COM port thread. If you're going to the trouble of create a device driver, you would want to base it on IOKit because of that likely IOServiceMatching query. You can find the Apple16X50Serial project mentioned in that post at the top of Apple's open source code list (go to the main page and pick an older OS release if you want to target something pre-10.6).
If your app is most useful with realtime data (e.g. the RouteBuddy app mentioned in the Python mailinglist thread can log current positions) then you will want to fetch updates from your web sources (hopefully they support long-polling) and convert them to basic NMEA RMC sentences. You do not want to do this from inside your driver code. Instead, divide your work up into kernel-land and user-land pieces that can communicate, and put as little of the code as possible into the kernel part.
If you want to let apps both read and write to these web services, your best bet would probably be to simulate a Garmin device. Garmin has more-or-less documented their protocol in the IntfSpec.pdf file included with their Device Interface SDK. Again, you'd want to split as much as you could into user-space code.
I was unable to find a project or utility that implements the kernel side of an IOKit-based virtual serial interface, but I'd be surprised if there wasn't one hiding somewhere out there. Unfortunately, most of the answers I found to that question were like this, with the developer being told to get busy writing a kext.
I'm not exactly sure how to accomplish what you're asking, but I may be able to lend some insight as to how you might begin to get it done. So here goes:
A GPS device shows up to most systems as nothing more than a serial device -- a.k.a. a COM port if you're dealing with Windows, /dev/ttySx if you're in *nix. By definition, a serial port's specific duty is to stream data across a bus, one block at a time. So, it would then follow logically that if you want to emulate the presence of a GPS device, you should gather the data you're consuming and put it into a stream that somehow acts like an active serial port.
There are, however, some complications you might want to consider:
Most GPS devices don't just send out location data; there's also information on satellite locations, fix quality, bearing, and so on. Then again, nobody's made any rules saying you have to make all that data available. There's probably more to this, but I'll admit that I need to do more research in this area myself.
I'm not sure how fast you can receive data when dealing with Google Latitude, etc., but any delays in receiving would definitely result in visible pauses in your "serial port"'s data stream. Again, this may not be as big a complication as it seems, because GPS devices are known to "burst" data across the bus anyway, but I'd definitely keep an eye on that. You want to make sure there's always a surplus of data coming across, not a shortage.
Along the way you'll also have to transform the coordinates you receive into valid GPS sentences, as well. You can find specifications for those, but I would definitely make friends with the NMEA standard -- even though it is a flawed standard, it's the one everyone seems to agree on anyway.
Hope this helped you, at least a little bit. Are there anymore details specific to your problem that you think could be useful in answering this question?
Take a look to Franson GPS Gate which allows you to connect to Google Earth among other things (like simulating GPS and so on). Is windows only though but I think you could get some useful ideas from it.
I haven't looked into it very much, but have you considered using Skyhook's SDK? It might provide you with some of what you are looking for. It's available for every major desktop and mobile OS.

Resources