arXiv subjects

Patrick Palmer

Publications and source records attributed to Patrick Palmer.

16 recordsLinked to original sources

arXiv2024

LEROjD: Lidar Extended Radar-Only Object Detection

Accurate 3D object detection is vital for automated driving. While lidar sensors are well suited for this task, they are expensive and have limitations in adverse weather conditions. 3+1D imaging radar sensors offer a cost-effective, robust alternative but face challenges due to their low resolution and high measurement noise. Existing 3+1D imaging radar datasets include radar and lidar data, enabling cross-modal model improvements. Although lidar should not be used during inference, it can aid the training of radar-only object detectors. We explore two strategies to transfer knowledge from the lidar to the radar domain and radar-only object detectors: 1. multi-stage training with sequential lidar point cloud thin-out, and 2. cross-modal knowledge distillation. In the multi-stage process, three thin-out methods are examined. Our results show significant performance gains of up to 4.2 percentage points in mean Average Precision with multi-stage training and up to 3.9 percentage points with knowledge distillation by initializing the student with the teacher's weights. The main benefit of these approaches is their applicability to other 3D object detection networks without altering their architecture, as we show by analyzing it on two different object detectors. Our code is available at https://github.com/rst-tu-dortmund/lerojd

cs.CV↗

arXiv2024

Multi-Object Tracking based on Imaging Radar 3D Object Detection

Effective tracking of surrounding traffic participants allows for an accurate state estimation as a necessary ingredient for prediction of future behavior and therefore adequate planning of the ego vehicle trajectory. One approach for detecting and tracking surrounding traffic participants is the combination of a learning based object detector with a classical tracking algorithm. Learning based object detectors have been shown to work adequately on lidar and camera data, while learning based object detectors using standard radar data input have proven to be inferior. Recently, with the improvements to radar sensor technology in the form of imaging radars, the object detection performance on radar was greatly improved but is still limited compared to lidar sensors due to the sparsity of the radar point cloud. This presents a unique challenge for the task of multi-object tracking. The tracking algorithm must overcome the limited detection quality while generating consistent tracks. To this end, a comparison between different multi-object tracking methods on imaging radar data is required to investigate its potential for downstream tasks. The work at hand compares multiple approaches and analyzes their limitations when applied to imaging radar data. Furthermore, enhancements to the presented approaches in the form of probabilistic association algorithms are considered for this task.

cs.RO↗

arXiv2023

Ego-Motion Estimation and Dynamic Motion Separation from 3D Point Clouds for Accumulating Data and Improving 3D Object Detection

New 3+1D high-resolution radar sensors are gaining importance for 3D object detection in the automotive domain due to their relative affordability and improved detection compared to classic low-resolution radar sensors. One limitation of high-resolution radar sensors, compared to lidar sensors, is the sparsity of the generated point cloud. This sparsity could be partially overcome by accumulating radar point clouds of subsequent time steps. This contribution analyzes limitations of accumulating radar point clouds on the View-of-Delft dataset. By employing different ego-motion estimation approaches, the dataset's inherent constraints, and possible solutions are analyzed. Additionally, a learning-based instance motion estimation approach is deployed to investigate the influence of dynamic motion on the accumulated point cloud for object detection. Experiments document an improved object detection performance by applying an ego-motion estimation and dynamic motion correction approach.

cs.RO↗

arXiv2023

Reviewing 3D Object Detectors in the Context of High-Resolution 3+1D Radar

Recent developments and the beginning market introduction of high-resolution imaging 4D (3+1D) radar sensors have initialized deep learning-based radar perception research. We investigate deep learning-based models operating on radar point clouds for 3D object detection. 3D object detection on lidar point cloud data is a mature area of 3D vision. Many different architectures have been proposed, each with strengths and weaknesses. Due to similarities between 3D lidar point clouds and 3+1D radar point clouds, those existing 3D object detectors are a natural basis to start deep learning-based 3D object detection on radar data. Thus, the first step is to analyze the detection performance of the existing models on the new data modality and evaluate them in depth. In order to apply existing 3D point cloud object detectors developed for lidar point clouds to the radar domain, they need to be adapted first. While some detectors, such as PointPillars, have already been adapted to be applicable to radar data, we have adapted others, e.g., Voxel R-CNN, SECOND, PointRCNN, and PV-RCNN. To this end, we conduct a cross-model validation (evaluating a set of models on one particular data set) as well as a cross-data set validation (evaluating all models in the model set on several data sets). The high-resolution radar data used are the View-of-Delft and Astyx data sets. Finally, we evaluate several adaptations of the models and their training procedures. We also discuss major factors influencing the detection performance on radar data and propose possible solutions indicating potential future research avenues.

cs.RO↗

arXiv2011

Microwave Observations of Edge-on Protoplanetary Disks: Program Overview and First Results

We are undertaking a multi-frequency Expanded Very Large Array (EVLA) survey of edge-on protoplanetary disks to probe the growth of solids in each disk, sedimentation of such material into the disk midplane, and the connection of these phenomena to the planet formation process. The projection of edge-on disk systems along our line of sight enables a study of the vertical stratification of large grains with fewer model dependencies than would be required for disks that are more face-on. Robust studies of the spatial distribution of grains up to ~1 cm in size are possible with the wavelength range and sensitivity of the EVLA. In this contribution we describe target selection and observational strategies. First results concerning the Class 0 source IRAS04368+2557 (L1527 IRS) are presented, including a study of this source's 8.46 GHz continuum variability over short and long time baselines and an indication that its protoplanetary disk may have a dearth of pebble-sized grains.

astro-ph.SR↗

arXiv2008

The Distribution, Excitation and Formation of Cometary Molecules: Methanol, Methyl Cyanide and Ethylene Glycol

We present an interferometric and single dish study of small organic species toward Comets C/1995 O1 (Hale-Bopp) and C/2002 T7 (LINEAR) using the BIMA interferometer at 3 mm and the ARO 12m telescope at 2 mm. For Comet Hale-Bopp, both the single-dish and interferometer observations of CH3OH indicate an excitation temperature of 105+/-5 K and an average production rate ratio Q(CH3OH)/Q(H2O)~1.3% at ~1 AU. Additionally, the aperture synthesis observations of CH3OH suggest a distribution well described by a spherical outflow and no evidence of significant extended emission. Single-dish observations of CH3CN in Comet Hale-Bopp indicate an excitation temperature of 200+/-10 K and a production rate ratio of Q(CH3CN)/Q(H2O)~0.017% at ~1 AU. The non-detection of a previously claimed transition of cometary (CH2OH)2 toward Comet Hale-Bopp with the 12m telescope indicates a compact distribution of emission, D<9'' (<8500 km). For the single-dish observations of Comet T7 LINEAR, we find an excitation temperature of CH3OH of 35+/-5 K and a CH3OH production rate ratio of Q(CH3OH)/Q(H2O)~1.5% at ~0.3 AU. Our data support current chemical models that CH3OH, CH3CN and (CH2OH)2 are parent nuclear species distributed into the coma via direct sublimation off cometary ices from the nucleus with no evidence of significant production in the outer coma.

astro-ph↗

arXiv2008

Gas and Dust Associated with the Strange, Isolated, Star BP Piscium

We have carried out a multiwavelength observational campaign demonstrating some of the remarkable properties of the infrared-bright variable star BP Psc. Surrounded by a compact dusty, gaseous disk, this little-studied late-G (or early-K) type star emits about 75% of its detected energy flux at infrared wavelengths. Evidence for accretion of gas in conjunction with narrow bi-polar jets and Herbig-Haro objects is apparently consistent with classification of BP Psc as a pre-main sequence star, as postulated in most previous studies. If young, then BP Psc would be one of the nearest and oldest known classical T Tauri stars. However, such an evolutionary classification encounters various problems that are absent or much less severe if BP Psc is instead a luminosity class III post-main sequence star. In this case, it would be the first known example of a first ascent giant surrounded by a massive molecular disk with accompanying rapid gas accretion and prominent jets and HH objects. In this model, the genesis of the massive dusty gaseous disk could be a consequence of the envelopment of a low mass companion star. Properties in the disk may be conducive to the current formation of planets, a gigayear or more after the formation of BP Psc itself.

astro-ph↗

arXiv2008

Non-thermal Emission in Sagittarius B?

We summarize three recent publications which suggest that the Galactic center region Sagittarius B (Sgr B) may contain non-thermal radio components (Crocker et al. 2007, Hollis et al. 2007 and Yusef-Zadeh et al. 2007a). Based on new VLA matched-resolution continuum data at 327 MHz and 1.4 GHz, we find no evidence for large scale non-thermal radio emission at these frequencies; the spectral behavior is likely determined by the complex summation of multiple HII region components with a wide range of emission measures and hence radio turn-over frequencies. Also, we discuss a possible additional interpretation of the radio continuum spectrum of individual component Sgr B2-F carried out by Yusef-Zadeh et al; confusion from nearby HII components with widely different turn-over frequencies may contribute to the the change in slope of the radio continuum in this direction at low frequencies. Finally, we discuss the uncertainties in the determination of the spectral index of the GBT continuum data of Sgr B carried out by Hollis et al. We find that the apparent spectral index determined by their procedure is also likely due to a summation over the many diverse thermal components in this direction.

astro-ph↗

arXiv2006

The Formaldehyde Masers in Sgr B2: Very Long Baseline Array and Very Large Array Observations

Observations of two of the formaldehyde (H2CO) masers (A and D) in Sgr B2 using the VLBA+Y27 (resolution ~0.01") and the VLA (resolution ~9") are presented. The VLBA observations show compact sources (<10 milliarcseconds, <80 AU) with brightness temperatures >10^8 K. The maser sources are partially resolved in the VLBA observations. The flux densities in the VLBA observations are about 1/2 those of the VLA; and, the linewidths are about 2/3 of the VLA values. The applicability of a core-halo model for the emission distribution is demonstrated. Comparison with earlier H2CO absorption observations and with ammonia (NH3) observations suggests that H2CO masers form in shocked gas. Comparison of the integrated flux densities in current VLA observations with those in previous observations indicates that (1) most of the masers have varied in the past 20 years, and (2) intensity variations are typically less than a factor of two compared to the 20-year mean. No significant linear or circular polarization is detected with either instrument.

astro-ph↗

arXiv2006

A Bima Array Survey of Molecules in Comets Linear (C/2002 T7) and Neat (C/2001 Q4)

We present an interferometric search for large molecules, including methanol, methyl cyanide, ethyl cyanide, ethanol, and methyl formate in comets LINEAR (C/2002 T7) and NEAT (C/2001 Q4) with the Berkeley-Illinois-Maryland Association (BIMA) array. In addition, we also searched for transitions of the simpler molecules CS, SiO, HNC, HN13C and 13CO . We detected transitions of methanol and CS around Comet LINEAR and one transition of methanol around Comet NEAT within a synthesized beam of ~20''. We calculated the total column density and production rate of each molecular species using the variable temperature and outflow velocity (VTOV) model described by Friedel et al.(2005).Considering the molecular production rate ratios with respect to water, Comet T7 LINEAR is more similar to Comet Hale-Bopp while Comet Q4 NEAT is more similar to Comet Hyakutake. It is unclear, however, due to such a small sample size, whether there is a clear distinction between a Hale-Bopp and Hyakutake class of comet or whether comets have a continuous range of molecular production rate ratios.

astro-ph↗

arXiv2005

BIMA Array Detections of HCN in Comets LINEAR (C/2002 T7) and NEAT (C/2001 Q4)

We present interferometric detections of HCN in comets LINEAR (C/2002 T7) and NEAT (C/2001 Q4) with the Berkeley-Illinois-Maryland Association (BIMA) Array. With a 25.4"x20.3" synthesized beam around Comet LINEAR and using a variable temperature and outflow velocity (VTOV) model, we found an HCN column density of N_T=6.4+-2.1x10^12 cm^-2, and a production rate of Q(HCN)=6.5+-2.2x10^26 s^-1, giving a production rate ratio of HCN relative to H_2O of ~3.3+-1.1x10^-3 and relative to CN of ~4.6+-1.5. With a 21.3"x17.5" synthesized beam around Comet NEAT and using a VTOV model, we found an HCN column density of N_T=8.5+-4.5x10^11 cm^-2, and a production rate of Q(HCN)=8.9+-4.7x10^25 s^-1, giving a production rate ratio of HCN relative to H_2O of ~7.4+-3.9x10^-4 and relative to CN of ~0.3+-0.2. For both comets, the production rates relative to H_2O are similar to those found in previous comet observations. For Comet LINEAR the production rate relative to CN is consistent with HCN being the primary parent species of CN, while for Comet NEAT it is too low for this to be the case.

astro-ph↗

arXiv2005

VLA observations of broad 6-cm excited state OH lines in W49A

Using the Very Large Array (VLA), we observed all three of the 6-cm lines of the doublet Pi 1/2, J=1/2 state of OH with sub-arcsecond resolution (about 0.4 arcsec) in W49A. While the spatial distribution and the range in velocities of the 6-cm lines are similar to those of the ground state (18-cm) OH lines, a large fraction of the total emission in all three 6-cm lines has large linewidths (about 5 -- 10 km/s) and is spatially-extended, very unlike typical ground state OH masers which typically are point-like at VLA resolutions and have linewidths less than 1 km/s. We find brightness temperatures of 5900 K, 4700 K, and greater than 730 K for the 4660-MHz, 4750-MHz, and 4765-MHz lines, respectively. We conclude that these are indeed maser lines. However, the gains are about 0.3, again very unlike the 18-cm lines which have gains greater than 10000. We compare the excited state OH emission with that from other molecules observed with comparable angular resolution to estimate physical conditions in the regions emitting the peculiar, low-gain maser lines. We also comment on the relationship with the 18-cm masers

astro-ph↗

arXiv2003

VLA observations of 6-cm excited OH

The VLA was used to determine precise positions for 4765-MHz OH maser emission sources toward star-forming regions which had been observed about seven months earlier with the Effelsberg 100-meter telescope. The observations were successful for K3-50, DR21EX, W75N, and W49A. No line was detected toward S255: this line had decreased to less than 5 per cent of the flux density observed only seven months earlier. The time-variability of the observed features during the past 30 years is summarised. In addition, to compare with the Effelsberg observations, the 4750-MHz and 4660-MHz lines were observed in W49A. These lines were found to originate primarily from an extended region which is distinguished as an exceptional collection of compact continuum components as well as by being the dynamical centre of the very powerful H_2 O outflow.

astro-ph↗

arXiv2003

Phase-Referenced VLBA Observations of OH Masers at 4765 MHz

We report VLBA observations of maser emission from the rotationally excited doublet Pi 1/2, J=1/2 state of OH at 4765 MHz. We made phase-referenced observations of W3(OH) at both 4765 MHz and 1720 MHz and found emission in three fields within a about 2000 AU diameter region and verified that in two of the three fields, 4765 MHz and 1720 MHz emission arise from the same position to within about 4 mas (about 5 AU diameter emission regions along an approximately N-S arc with linear extent about 500 AU. In addition, we carried out phase-referenced observations of 4765 MHz emission from K3-50. We searched for the 4765 MHz line in W49 (without phase referencing) and W75N (phase-referenced to the strongest 4765 MHz maser feature in DR21EX); we were unable to detect these sources with the VLBA. For 2 1/2 years (including the dates of the VLBA observations), we carried out monitoring observations of 4765 MHz emission with the VLA. Constraints on models for maser emission at 1720 MHz and 4765 MHz are derived from the observations. These observations are then briefly compared with existing models.

astro-ph↗

arXiv2003

The Formaldehyde Masers in NGC 7538 and G29.96-0.02: VLBA, MERLIN, and VLA Observations

The 6 cm formaldehyde (H2CO) maser sources in the compact HII regions NGC 7538-IRS1 and G29.96-0.02 have been imaged at high resolution (beams < 50 mas). Using the VLBA and MERLIN, we find the angular sizes of the NGC 7538 masers to be ~10 mas (30 AU) corresponding to brightness temperatures ~10^8 K. The angular sizes of the G29.96-0.02 masers are ~20 mas (130 AU) corresponding to brightness temperatures ~10^7 K. Using the VLA, we detect 2 cm formaldehyde absorption from the maser regions. We detect no emission in the 2 cm line, indicating the lack of a 2 cm maser and placing limits on the 6 cm excitation process. We find that both NGC 7538 maser components show an increase in intensity on 5-10 year timescales while the G29.96-0.02 masers show no variability over 2 years. A search for polarization provides 3-sigma upper limits of 1% circularly polarized and 10% linearly polarized emission in NGC 7538 and of 15% circularly polarized emission in G29.96-0.02. A pronounced velocity gradient of 28 km/s/arcsecond (1900 km/s/pc) is detected in the NGC 7538 maser gas.

astro-ph↗

arXiv1996

Nomenclature of The Galactic Center Radio Sources

Since its identification as the Galactic Center, the radio source Sagittarius A has been a target of intense research. Due to the high density of sources in the Galactic Center and differing observing techniques, the nomenclature of sources in this region has changed over the years, with sources having several names, as well as the same names being used for different sources. We review this historical evolution in the context of current and previous scientific discussions, and outline how, why and when some of the commonly accepted names of Galactic Center sources were established.

astro-ph↗