Collision-Free Path Planning and Delivery Sequence Optimization in Noncoplanar Radiation Therapy

LENAS: Learning-Based Neural Architecture Search and Ensemble for 3-D Radiotherapy Dose Prediction

Radiation therapy treatment planning requires balancing the delivery of the target dose while sparing normal tissues, making it a complex process. To streamline the planning process and enhance its quality, there is a growing demand for knowledge-based planning (KBP). Ensemble learning has shown impressive power in various deep learning tasks, and it has great potential to improve the performance of KBP. However, the effectiveness of ensemble learning heavily depends on the diversity and individual accuracy of the base learners. Moreover, the complexity of model ensembles is a major concern, as it requires maintaining multiple models during inference, leading to increased computational cost and storage overhead. In this study, we propose a novel learning-based ensemble approach named LENAS, which integrates neural architecture search with knowledge distillation for 3-D radiotherapy dose prediction. Our approach starts by exhaustively searching each block from an enormous architecture space to identify multiple architectures that exhibit promising performance and significant diversity. To mitigate the complexity introduced by the model ensemble, we adopt the teacher-student paradigm, leveraging the diverse outputs from multiple learned networks as supervisory signals to guide the training of the student network. Furthermore, to preserve high-level semantic information, we design a hybrid loss to optimize the student network, enabling it to recover the knowledge embedded within the teacher networks. The proposed method has been evaluated on two public datasets: 1) OpenKBP and 2) AIMIS. Extensive experimental results demonstrate the effectiveness of our method and its superior performance to the state-of-the-art methods. Code: github.com/hust-linyi/LENAS.

A Cooperative Hunting Method for Multi-USV Based on the A* Algorithm in an Environment with Obstacles

A single unmanned surface combatant (USV) has poor mission execution capability, so the cooperation of multiple unmanned surface ships is widely used. Cooperative hunting is an important aspect of multi USV collaborative research. Therefore, this paper proposed a cooperative hunting method for multi-USV based on the A* algorithm in an environment with obstacles. First, based on the traditional A* algorithm, a path smoothing method based on USV minimum turning radius is proposed. At the same time, the post order traversal recursive algorithm in the binary tree method is used to replace the enumeration algorithm to obtain the optimal path, which improves the efficiency of the A* algorithm. Second, a biomimetic multi USV swarm collaborative hunting method is proposed. Multiple USV clusters simulate the hunting strategy of lions to pre-form on the target's path, so multiple USV clusters do not require manual formation. During the hunting process, the formation of multiple USV groups is adjusted to limit the movement and turning of the target, thereby reducing the range of activity of the target and improving the effectiveness of the algorithm. To verify the effectiveness of the algorithm, two sets of simulation experiments were conducted. The results show that the algorithm has good performance in path planning and target search.

Navigation of a Fuzzy-Controlled Wheeled Robot Through the Combination of Expert Knowledge and Data-Driven Multiobjective Evolutionary Learning

This article proposes a navigation scheme for a wheeled robot in unknown environments. The navigation scheme consists of obstacle boundary following (OBF), target seeking (TS), and vertex point seeking (VPS) behaviors and a behavior supervisor. The OBF behavior is achieved by a fuzzy controller (FC). This article formulates the FC design problem as a new constrained multiobjective optimization problem and finds a set of nondominated FC solutions through the combination of expert knowledge and data-driven multiobjective ant colony optimization. The TS behavior is achieved by new fuzzy proportional-integral-derivative (PID) and proportional-derivative (PD) controllers that control the orientation and speed of the robot, respectively. The VPS behavior is proposed to shorten the navigation route by controlling the robot to move toward a new subgoal determined from the vertex point of an obstacle. A new behavior supervisor that manages the switching among the OBF, TS, and VPS behaviors in unknown environments is proposed. In the navigation of a real robot, a new robot localization method through the fusion of encoders and an infrared localization sensor using a particle filter is proposed. Finally, this article presents simulations and experiments to verify the feasibility and advantages of the navigation scheme.

Automated Segmentation of the Clinical Target Volume in the Planning CT for Breast Cancer Using Deep Neural Networks

3-D radiotherapy is an effective treatment modality for breast cancer. In 3-D radiotherapy, delineation of the clinical target volume (CTV) is an essential step in the establishment of treatment plans. However, manual delineation is subjective and time consuming. In this study, we propose an automated segmentation model based on deep neural networks for the breast cancer CTV in planning computed tomography (CT). Our model is composed of three stages that work in a cascade manner, making it applicable to real-world scenarios. The first stage determines which slices contain CTVs, as not all CT slices include breast lesions. The second stage detects the region of the human body in an entire CT slice, eliminating boundary areas, which may have side effects for the segmentation of the CTV. The third stage delineates the CTV. To permit the network to focus on the breast mass in the slice, a novel dynamically strided convolution operation, which shows better performance than standard convolution, is proposed. To train and evaluate the model, a large dataset containing 455 cases and 50 425 CT slices is constructed. The proposed model achieves an average dice similarity coefficient (DSC) of 0.802 and 0.801 for right-0 and left-sided breast, respectively. Our method shows superior performance to that of previous state-of-the-art approaches.

FC-RRT*: An Improved Path Planning Algorithm for UAV in 3D Complex Environment

In complex environments, path planning is the key for unmanned aerial vehicles (UAVs) to perform military missions autonomously. This paper proposes a novel algorithm called flight cost-based Rapidly-exploring Random Tree star (FC-RRT*) extending the standard Rapidly-exploring Random Tree star (RRT*) to deal with the safety requirements and flight constraints of UAVs in a complex 3D environment. First, a flight cost function that includes threat strength and path length was designed to comprehensively evaluate the connection between two path nodes. Second, in order to solve the UAV path planning problem from the front-end, the flight cost function and flight constraints were used to inspire the expansion of new nodes. Third, the designed cost function was used to guide the update of the parent node to allow the algorithm to consider both the threat and the length of the path when generating the path. The simulation and comparison results show that FC-RRT* effectively overcomes the shortcomings of standard RRT*. FC-RRT* is able to plan an optimal path that significantly improves path safety as well as maintains has the shortest distance while satisfying flight constraints in the complex environment. This paper has application value in UAV 3D global path planning.

Patient-Specific Collision Zones for 4π Trajectory Optimized Radiation Therapy

PURPOSE

The 4π methodology determines optimized non-coplanar sub arcs for stereotactic radiation therapy which minimize dose to organs-at-risk. Every combination of treatment angle is examined, but some angles are not appropriate as a collision would occur between the gantry and the couch or the gantry and the patient. Those combinations of couch and gantry angles are referred to as collision zones. A major barrier to applying 4π to stereotactic body radiation therapy (SBRT) is the unknown shape of the collision zones, which are significant as patients take up a large volume within the 4π sphere. This study presents a system which determines patient-specific collision zones, without additional clinical steps, to enable safe and deliverable non-coplanar treatment trajectories for SBRT patients.

METHODS

To augment patient's computed tomography (CT) scan, full body scans of patients in treatment position were acquired using an optical scanner. A library of a priori scans (N = 25) was created. Based on the patients treatment position and their body dimensions, a library scan is selected and registered to the CT scan of the patient. Next, a model of the couch and immobilization equipment is added to the patient model. This results in a patient model that is then aligned with a model of the treatment linac in a "virtual treatment room", where both components can be rotated to test for collisions. To test the collision detection algorithm, an end-to-end test was performed using a cranial phantom. The registration algorithm was tested by comparing the registered patient collision zones to those generated by using the patient's matching scan.

RESULTS

The collision detection algorithm was found to have a 97.80% accuracy, a 99.99% sensitivity and a 99.99% negative predictive value (NPV). Analysis of the registration algorithm determined that a 6 cm buffer was required to achieve a 99.65% mean sensitivity, where a sensitivity of unity is considered to be a requirement for safe treatment delivery. With a 6 cm buffer the mean accuracy was 86.70% and the mean NPV was 99.33%.

CONCLUSIONS

Our method of determining patient-specific collision zones can be accomplished with minimal user intervention based on an a priori library of body surface scans, thus enabling the safe application of 4π SBRT.

Solving Trajectory Optimization Problems in the Presence of Probabilistic Constraints

The objective of this paper is to present an approximation-based strategy for solving the problem of nonlinear trajectory optimization with the consideration of probabilistic constraints. The proposed method defines a smooth and differentiable function to replace probabilistic constraints by the deterministic ones, thereby converting the chance-constrained trajectory optimization model into a parametric nonlinear programming model. In addition, it is proved that the approximation function and the corresponding approximation set will converge to that of the original problem. Furthermore, the optimal solution of the approximated model is ensured to converge to the optimal solution of the original problem. Numerical results, obtained from a new chance-constrained space vehicle trajectory optimization model and a 3-D unmanned vehicle trajectory smoothing problem, verify the feasibility and effectiveness of the proposed approach. Comparative studies were also carried out to show the proposed design can yield good performance and outperform other typical chance-constrained optimization techniques investigated in this paper.