Ping Shut Talks Human-Computer Interaction of VR-VR System - Action Continued

Rift's Constellation

Inside Rift's accessories there is an accessory that looks like a desk microphone. This device is not something that lets you talk to it. Many reviewers say it is a camera, but how did the camera not see the camera? However, it is really a camera. It has a light filter outside the lens, so you can't see the camera. This accessory is the core part of the Rift – motion capture infrared camera.

Rift uses a dual positioning system, one is an infrared camera optical positioning system, and the other is a nine degree of freedom inertial positioning system. Infrared camera can identify the LED lights on the helmet. The camera can obtain an LED image 10ms. The spatial position of the helmet is obtained by calculating the difference of the LED points (the algorithm is called PnP). Students who need specific algorithm principles can refer to this article.

Decrypting Oculus Rift's Positioning Technology

The advantage of the infrared positioning system is that it is simple to install, the delay is small, there is no accumulative error; The disadvantage is the general precision, it is invalid after covering. The solution to occlusion is to deploy a very large number of infrared cameras around, which will incur too high costs. In order to balance the cost and reliability Oculus added another positioning technology in the helmet, a nine-axis positioning system.

What is the nine axis is actually composed of three sets of positioning information. The first group, the speed parameters, are the movement speeds of the object in the three directions of space X, Y, Z; the second group, the angular speed parameters, are the angular speeds of the object in space, the pitch angle, the left and right corners and the roll angle; Three groups, geomagnetic parameters, three angle parameters with respect to the geomagnetic field. Usually we use the first two parameters to know the trajectory of the object in space, but the inertial positioning system has one of the biggest problems is the cumulative error, the third group of parameters is used to reduce the error, this error can only be corrected, can not be done To eliminate. Combined with the absolute positioning method of infrared positioning system, this problem is well solved.

In summary, absolute positioning is used when the camera can capture the helmet LED light, and the inertial positioning system is used when the camera is blocked.

Vive's Lighthouse

The Vive uses laser positioning technology, which emits horizontal and vertical laser beams through two high-altitude lighthouses. The emitted beam is received by the Vive's helmet and handle. The information is used to calculate the spatial position of the helmet and the handle. attitude. The Vive's helmet and handle are covered with 70 light sensors.

Lighthouse should be said to be the most cost-effective and accurate positioning method among the three devices. Both laser technology and photosensitive sensor technology are very mature technologies. Due to mature technology and simple algorithm, there is almost no delay in interaction and the response is rapid. This may be the reason why Vive started R&D but it was even earlier than Rift. The follow-up manufacturers are likely to be cottages. The only disadvantage of this method is that the equipment has a certain space requirement for layout, which cannot be promoted for places like Internet cafes.

PS VR Visible Light Positioning System

The principle of visible light positioning is similar to human eyes, PS VR has dual cameras, dual cameras simultaneously view, you can get the image of the object separately, due to the difference in the position of the two cameras, two photos will have some differences, because the camera between The distance is known so that the position information and posture of the object can be calculated. As shown, the recognition objects are a blue cross on the helmet, green and red balls on the handle.

The benefits of visible light are the low cost, simple algorithm, easy installation, and high stability. The disadvantage is that the positioning accuracy is not high, it cannot be positioned after shielding, and it is easily affected by ambient light. The low-cost visible light technology Ping Shu thinks that early VR products may use this type of method for reference. It only takes half the price to have a slightly different experience. The cost advantage is so obvious.

Although the input devices of the three major manufacturers are different from each other in terms of technology implementation, they are similar in terms of large features. Both are identified by the handle and the optical methods are used. However, this kind of interaction is still not natural enough and not direct enough. The best way to interact is to not use it, not to get it, and to use human body language to interact. The current progress in this area is also very rapid. Representatives are Microsoft's Kinect and Leap Motion. Oculus has already deployed early on in this area and acquired the Israeli gesture recognition company Pebbles Interfaces as early as 15 years ago. It is foreseeable that gesture recognition in the future will certainly be an important means of VR interaction.

One of the arguments has been chattering. Does the user really like to hold the handle or not? You see who you support.