CRPC ® (Cognitive Radio Protocol Cracking), namely, cognitive radio protocol reverse parsing technology, is independently developed by Beijing Lizheng Technology and is the first to be applied to the detection and defense of low altitude UAVs. It uses radio spectrum sensing technology (similar to the feature extraction of viruses by Internet anti-virus software) to extract UAV communication signals (mainly remote control and image transmission signals) from background noise and conduct in-depth reverse analysis. The UAV communication protocol is analyzed offline and features (electronic fingerprint ID) are extracted, including but not limited to channel distribution, modulation scheme, coding scheme, pilot sequence, synchronization sequence, link layer structure, etc. Based on the above analysis results, a UAV detection database is formed, and real-time detection and matching are performed in the actual radio environment. Finally, features are applied to real-time detection (fast matching and recognition of features) and precision strike (distinguishing different UAVs to strike) of UAVs.
RFML (Radio Frequency Machine Learning), that is, wireless signal machine learning technology, widely extracts signal features through neural networks to achieve the recognition of UAV signals, IDs and other functions. In the recognition and defense of low altitude aircraft, its neural network does not depend on the special design of the transmitted signal. The range of signal features that can be extracted and used is longer, and the learning and training time is shorter. It can quickly identify different types of low altitude aircraft, so as to realize the signal learning of unknown unmanned aircraft and the rapid expansion and improvement of the aircraft type library. Compared with traditional UAV detection and recognition technology, RFML has the following advantages:
Quickly support unknown UAVs For the support of DIY, traversing machine and other new UAVs, only a certain amount of UAV signals need to be obtained, and then through the training of neural network, the support for this new UAV can be achieved. Compared with traditional manual signal feature extraction, the training time of neural network has significant advantages such as short time and less manual participation Wider range of radio applications
The traditional UAV detection technology mostly depends on the special design of the transmitted signal, and it is difficult to use other unique features of the signal (such as CFO). However, these features can be extracted and utilized by RFML using neural networks. In other words, the range of signal features that RFML can use is far greater than that of traditional technologies. Therefore, the range that RFML can be applied is far greater than that of traditional technologies.
Both TDOA and AOA technologies need high-precision precision detectors as the basis, otherwise the positioning technology will lose its significance. Therefore, with the help of high-precision UAV detection technology, TDOA/AOA technology will realize that any UAV can not only "detect", but also "follow accurately".
For this reason, Lizheng Technology took the lead in introducing the "TDOA+(Time Difference of Arrival Plus)" UAV precise identification and positioning technology, integrating TDOA and CRPC technology, realizing the triple functions of "precise identification", "precise positioning" and "precise strike" of the target UAV, and greatly improving the system's practical application performance.

Technical features: * Passive detection: It only receives passively without transmitting any electromagnetic signal, and can detect the UAV model and its electronic fingerprint ID; * Precise identification: it can accurately identify different UAVs of the same brand, model and frequency band;
* Precise positioning: it can accurately measure the position of UAV/remote controller and indicate the direction; * Track tracking: It has the function of multi-target track tracking, and can display the flight track of multiple UAVs in different colors when they are detected at the same time; * Black and white list: the black and white list can be marked with one key, and the white list drone will not alarm; * Full model library: support the whole series of DJI, market mainstream brands and DIY UAVs covering more than 98% of the models on the market;