We present a generalized mapping framework that can withstand the challenges incurred by working in unstructured outdoor environments, such as a snowy forest. The proposed method takes advantage of a sensor fusion scheme, where sensors such as cameras and lidars are used in order to reconstruct the surrounding natural environment. Although mapping techniques such as SLAM and ICP cannot themselves properly handle the complexity of natural scenes, they do have the potential to contribute to the global solution in a proposed sensor fusion scheme, based on a factor graph architecture. In this paper, we propose an innovative map registration scheme for visual maps, and show how it can improve the reconstruction quality after data fusion. We also analyze the behavior and sensitivity of factor graphs to uncertainties, by comparing the residual error with different parameter combinations such as variances, using an exhaustive grid search with ground truth comparison. Finally, we suggest an ICP-inferred loop closure, capable of compensating position and attitude drift. The experiments are carried out by recording in a snowy forest using a wearable sensor suite. In the experiments, ground truth was acquired using a millimeter-accurate total station. The proposed framework is shown to be robust and likewise capable of providing estimates that are otherwise unattainable using classic techniques, such as visual SLAM and ICP for lasers. Finally, a visible improvement in the map reconstruction quality is shown, and the proposed framework achieves a translation error of 0.36 m.

The inspection and maintenance of large-scale industrial structures are major challenges that require time-efficient and reliable solutions to ensure the healthy condition of structures during operation. Autonomous robots may provide a promising solution for this purpose. In particular, they could lead to faster and more reliable inspection and maintenance without direct intervention from human operators. In this paper, we present a custom magnetic crawler system, and sensor suit and sensing modalities to enable such robotic operation. We first describe a localization framework based on a mesh created from a point cloud coupled with Inertial Measurement Unit (IMU) and Ultra-Wide Band (UWB) readings. Next, we introduce a mapping framework that relies on a 3D laser, and explicitly state how autonomous navigation and obstacle avoidance can be developed. Lastly, we present how ultrasonic guided waves (UGWs) are integrated into the system to provide accurate robot localization and structural feature mapping by relying on acoustic reflections in combination with the other systems. It is envisioned that long-range inspection capabilities that are not yet available in current industrial mobile platforms could emerge from the designed robotic system.

Breast conservation rate is being increasingly used nowadays as a marker of breast cancer care among hospitals. Searching for the ideal technique to predict the feasibility of BCS is ongoing. For this matter, the preoperative MRIs of 169 patients operated with radical or conservative surgery were reviewed. We estimated the tumor volume (TV) and breast volume (BV) on enhanced 3D‐MRI and compared the tumor‐to‐breast volume ratio (TV/BV) in both groups. The mean ratio was 9.5% in the mastectomy group and 1.7% in the BCS group. A tumor‐to‐breast volume ratio less than 4% seemed to favor the adoption of a conservative option. Our data suggest that preoperative 3D‐MRI can orient the surgical approach by assessing the TV/BV ratio, increasing lumpectomy rates with clear margins and good cosmetic outcome.

Autonomous smart cars are an integral part of smart cities. This paper partly addresses the localization and mapping challenges that need to be addressed in order to autonomously control the movement of a car. Further, we explore multiple use-cases such as camera pose retrieval, thermal visual SLAM, and GPS-constrained Visual SLAM. Finally, we conduct our experiments using the classic and our own, modified version of ORB-SLAM, using both the KITTI Dataset and our own datasets, recorded using Valeo's autonomous car prototypes.