Needed for Augmented Reality the user needs a combination of display (Smartphone, HMD), tracking (A-GPS, accelerometers, gyros, optical sensors, RFID and onboard compasses), input devices (gloves, goggles, Smartphone) and software (image registration etc).
Originally Augmented Reality required special goggles or glasses that used a special array of mirrors to enable the user to see the real word with virtual information overlaid. But thanks to the widespread availability of Smartphone's and the increasing amount of Augmented Reality applications available to freely download you can also use the camera of your phone to engage in AR wherever you are, 24/7.
Now more recently the use of portable projectors can be used to overlay content into the real world (known as Spatial Displays) and although still in its infancy portable projection holds huge potential as the ability to provide passive haptic feedback through touching physical objects that have content projected onto them. Although currently the main AR display is the Smartphone due to its widespread appeal and its relatively wide market use.
Originally with the use of HMD as the earliest implementation of AR, as well as the HMD a tracker system needed to be attached and these systems would become quite bulky in size. But again thanks to the widespread availability of Smartphone's you can have AR in your pocket and all the possibilities to find you're nearest restaurant. Using A-GPS (assisted global positioning service), accelerometers and onboard compasses your can just point your camera phone at a restaurant to get the phone number to make a reservation or even to read reviews, if you even feel like it the you can have a look at the menu to decide if you like what you see. If you don't fancy anything on the menu then point your camera around you to look for similar restaurants and see what direction they are in.
As with the earlier uses of Augmented Reality an input device them would have needed gloves that could be tracked to display virtual objects. Currently input devices are also a research and development question as new tracking techniques could be employed with wand like devices for example.
As the whole idea of AR is to super impose virtual content into the real world, which is captured by a camera and displayed through the display device chosen. The visual content has to be able to be aligned with the captured image from the camera. Depending on whether these are fixed points or moving targets makes no difference, as the virtual content still needs to be registered and positioned in relation to the user, this is called 'image registration'. This is why highly developed software needs to be employed to make sure that this is possible, and works.