Dayat Kurniawan

Autonomous radar interference is a challenging problem in autonomous vehicle systems. Interference signals can decrease the signal-to-interference-noise ratio (SINR), and this condition decreases the performance detection of autonomous radar. This paper exploits a neural network and signal decomposition to detect and mitigate radar interference in autonomous vehicle applications. A neural network (NN) with four inputs, one hidden layer, and one output is trained with various signal-to-noise (SNR), interference radar bandwidth, and sweep time of autonomous radar. Four inputs of NN represent SNR, mean, total harmonic distortion (THD), and root means square (RMS) of the received radar signal. Variational mode decomposition (VMD) and zeroing based on a constant false alarm rate (CFAR-Z) are used to mitigate radar interference. VMD algorithm is applied to decompose interference signals into multi-frequency sub-band. As a result, the proposed neural network can detect radar interference, and NN-VMD-CFAR-Z can increase SINR up to 2dB higher than the NN-CFAR-Z algorithm.

Radar is one of the sensors that have significant attention to be implemented in an autonomous vehicle since its robustness under many possible environmental conditions such as fog, rain, and poor light. However, the implementation risks interference because of transmitting and/or receiving radar signals from/to other vehicles. This interference will increase the floor noise that can mask the target signal. This paper proposes multiplicative-adaptive filtering and Hilbert transform to mitigate the interference effect and maintain the target signal detectability. The method exploited the trade-off between the step-size and sidelobe effect on the least mean square-based adaptive filtering to improve the target detection accuracy, especially in the long-range case. The numerical analysis on the millimeter-wave frequency modulated continuous wave radar with multiple interferers concluded that the proposed method could maintain and enhance the target signal even if the target range is relatively far from the victim radar.

Time of Flight (ToF) Lidar range finder is widely used in various applications such as automotive driving, marine environment surveying, atmospheric observing, and physical distancing monitoring. The ToF Lidar offers a high-resolution range, which can be achieved by increasing the modulation frequency. However, when the modulation frequency is increased, the maximum unambiguous range tends to decrease. To overcome such a problem, we develop a novel ToF Lidar range finder based on dual-modulation frequencies switching, capable of providing an extended range while maintaining the high-resolution. In the design, the modulation frequency is alternately changed between 2 MHz and 10 MHz to modulate the intensity of the laser diode as a transmitter. We operate an avalanche photodetector as a receiver to obtain a reflection of modulated light from a target. By switching between dual-modulation frequencies, the photodetector detects two-phase shifts, which are calculated by our phase-shift algorithm. It is shown that the proposed algorithm can solve the ambiguous range problem. As proof of the concept, we experimentally demonstrate a range measurement of a target up to 55 m distance, where the ambiguous range is experienced four times if only a single modulation frequency of 10 MHz is used. Interestingly, the range ambiguity no longer appears after applying the proposed design and algorithm.

In this paper, 1 × 2 MIMO of Palm Tree Coplanar Vivaldi Antenna is presented that simulated at 0.5–4.5 GHz. Some GPR applications require wideband antennas starting from a frequency below 1 GHz to overcome high material loss and achieve deeper penetration. However, to boost the gain, antennas are set up in MIMO and this is costly due to the large size of the antenna. When configuring MIMO antenna in the E-plane, there is occasionally uncertainty over which antenna model may provide the optimum performance in terms of return loss, mutual coupling, directivity, beam squint, beam width, and surface current using a given substrate size. However, the configuration of E-plane antenna in MIMO has an issue of mutual coupling if the distance between elements is less than 0.5λ. Furthermore, it produces grating lobes at high frequencies.We implement several types of patch structures by incorporating the truncated, tilt shape, Hlbert and Koch Fractal, Exponential slot, Wave slot, the lens with elips, and metamaterial slot to the radiator by keeping the width of the substrate and the shape of the feeder. The return loss, mutual coupling, directivity, beam squint, beamwidth, and surface current of the antenna are compared for 1 × 2 MIMO CVA. A continuous patch MIMO has a spacing of 0.458λ at 0.5 GHz, which is equivalent to its element width. From the simulation, we found that Back Cut Palm Tree (BCPT) and Horizontale Wave Structure Palm Tree (HWSPT) got the best performance of return loss and mutual scattering at low-end frequency respectively. The improvement of directivity got for Metamaterial Lens Palm Tree (MLPT) of 4.453 dBi if compared with Regular Palm Tree-Coplanar Vivaldi Antena (RPT) at 4 GHz. Elips Lens Palm Tree (ELPT) has the best beam squint performance across all frequencies of 0°. It also gots the best beamwidth at 4.5 GHz of 3.320. In addition, we incorporate the MLPT into the radar application.

Collision is the main issue in safe transportation, including in the railway system. Sensor systems have been developed to detect obstacles to prevent a collision, such as using cameras. One disadvantage of the camera systems is that performance detection decreases in a not clean environment, like the target position behind the fogs. This paper discusses the development of frequency modulated continuous wave (FMCW) radar signal processing for high-speed railway collision avoidance. The development of radar signal processing combines a two-dimensional constant false alarm rate (2D-CFAR) and robust principal component analysis (RPCA) to detect moving targets under clutter. Cell average (CA) and Greatest of CA (GOCA) CFAR are evaluated under a cluttered wall environment along the railway track. From the experiment, the development of FMCW radar can detect stationary or moving obstacles around 675 meters in front of the locomotive. Combining 2D-CFAR and RPCA algorithm outperforms average background subtraction in extracting moving targets from strong clutter signals along the railway track.

A compact radar system becomes a necessity in order to build a lightweight navigation radar. Researches were conducted to create an integrated radar system in the form of one board of 9.3 GHz navigation radar system. Microstrip technology was used for its simple structure and miniaturization at high frequencies. Power divider as a part of the radar system was used to divide the signal from Voltage-Controlled Oscillator (VCO) to Transmitter front end and mixer on the Receiver end. This paper proposed the design and simulation of the Wilkinson Power Divider on Duroid RT5880 substrate using Advanced Design System (ADS) software. The S parameters were analyzed to see the performance of the divider. The proposed design already met the expected requirement of the power divider with more than -30 dB return loss, -3.059 dB insertion loss, and -29.541 dB isolation between divider output ports.

Direct-conversion is one of the receiver architecture units that converts of RF signal to IF signal into in-phase and quadrature (I/Q) form. One advantage of direct conversion is a simple architecture and cheaper than other receiver architecture. This architecture was implemented on FMCW radar for a navigation system which developed by P2ET-LIPI since 2016. One problem of direct conversion is I/Q imbalance. I/Q imbalance occurs when I/Q signal have different amplitude and phase. I/Q imbalance will produce image/mirror signal after FFT processing and it will be detected as a target of radar. The trigonometric function has been implemented to reducing I/Q imbalance by correcting amplitude and phase. This algorithm developed using Visual C# 2015 Community Edition. The algorithm consists of dc removal function, amplitude correction, and phase correction. The experiment shows that the algorithm has good performance at higher SNR to correcting I/Q imbalance.

Data acquisition and Signal Processing is an important part of frequency modulated continuous wave (FMCW) radar system. In recent year P2ET-LIPI was developing FMCW navigation radar system in portable size. The main component of data acquisition for navigation radar is highspeed Analog to Digital Conversion (ADC) USB-2020 board and signal processing algorithm has been developed using Visual C# 2015 Community Edition. USB-2020 board has two channel ADC on board with 12 bits resolution and operated in a burst-io mode to get maximum sampling rate at 20 MS/s. A timing generator for synchronization between ADC and linear frequency modulation (LFM) chirp generator is developed using Complex Programmable Logic Device (CPLD) Max II series based on VHSIC hardware description language (VHDL). Signal processing unit consists of windowing function to reduce side- lobe suppression, Fast Fourier Transform (FFT) function to get range information and adaptive threshold using three algorithms of Constant False Alarm Rate (CFAR): Cell Averaging (CA), Order Statistic (OS) and Variable Index (VI) CFAR. The experiment shows that the data acquisition module can capture In-phase and Quadrature (I/Q) beat signal and signal processing algorithm can detect object/target from the noise.

Target detection is a mandatory task of radar system so that the radar system performance is mainly determined by its detection rate. Constant False Alarm Rate (CFAR) is a detection algorithm commonly used in radar systems. This method is divided into several approaches which have different performance in the different environments. Therefore, this paper proposes an ensemble neural network based classifier with a variation of hidden neuron number for classifying the radar environments. The result of this research will support the improvement of the performance of the target detection on the radar systems by developing such an adaptive CFAR. Multi-layer perceptron network (MLPN) with a single hidden layer is employed as the structure of base classifiers. The first step of this research is the evaluation of the hidden neuron number giving the highest accuracy of classification and the simplicity of computation. According to the result of this step, the three best structures are selected to build an ensemble classifier. On the ensemble structure, all of those three MLPN outputs then be collected and voted for getting the majority result in order to decide the final classification. The three possible radar environments investigated are homogeneous, multiple-targets and clutter boundary. According to the simulation results, the ensemble MLPN provides a higher detection rate than the conventional single MLPNs. Moreover, in the multiple-target and clutter boundary environments, the proposed method is able to show its highest performance.

Pulse compression is a method for improving the range resolution of the radar systems. Linear Frequency Modulation (LFM) is a type of pulse compression which is commonly used in the radar applications because of its simplicity and Doppler tolerance. However this method provides relatively high first peak sidelobe level within around -13.3 dB. This paper presents a developed curve-shaped Nonlinear Frequency Modulation (NLFM) using second degree polynomial approach for radar pulse compression in order to reduce the sidelobe level. The radar signal is assumed has bandwidth 30 MHz while the length of pulse signal as long as 10e-6 s. The curve-shaped chirp waveform is formed by the polynomial equation utilizing the LFM waveform as the base line. According to the simulation results and analysis, using a certain polynomial equation, the curve-shaped NLFM with negative orientation provides lower sidelobe level than the LFM. The best first peak sidelobe level achieved by this proposed NLFM is approximately -18.6 dB.