Robust Path Planning via Learning from Demonstrations for Robotic Catheters in Deformable Environments

TL;DR

The proposed C-GAIL path planner outperforms the state-of-the-art GAIL approach and can provide enhanced support to the user in navigating the catheter towards the target with greater accuracy, effectively meeting clinical accuracy requirements.

Abstract

Objective: Navigation through tortuous and deformable vessels using catheters with limited steering capability underscores the need for reliable path planning. State-of-the-art path planners do not fully account for the deformable nature of the environment. Methods: This work proposes a robust path planner via a learning from demonstrations method, named Curriculum Generative Adversarial Imitation Learning (C-GAIL). This path planning framework takes into account the interaction between steerable catheters and vessel walls and the deformable property of vessels. Results: In-silico comparative experiments show that the proposed network achieves a 38% higher success rate in static environments and 17% higher in dynamic environments compared to a state-of-the-art approach based on GAIL. In-vitro validation experiments indicate that the path generated by the proposed C-GAIL path planner achieves a targeting error of 1.26$\pm$0.55mm and a tracking error of 5.18$\pm$3.48mm. These results represent improvements of 41% and 40% over the conventional centerline-following technique for targeting error and tracking error, respectively. Conclusion: The proposed C-GAIL path planner outperforms the state-of-the-art GAIL approach. The in-vitro validation experiments demonstrate that the path generated by the proposed C-GAIL path planner aligns better with the actual steering capability of the pneumatic artificial muscle-driven catheter utilized in this study. Therefore, the proposed approach can provide enhanced support to the user in navigating the catheter towards the target with greater accuracy, effectively meeting clinical accuracy requirements. Significance: The proposed path planning framework exhibits superior performance in managing uncertainty associated with vessel deformation, thereby resulting in lower tracking errors.

Figures (10)

• Figure 1: Parameterization of a robotic catheter agent: the catheter tip has configuration $\bm{q_t}$ at time $t$. The agent can perform an insertion movement $\Delta_l$ along the $y_A$ axis and can bend with angle $\alpha$ about the $x_A$ axis and with angle $\gamma$ about the $z_A$ axis, respectively, in the tip frame $\mathcal{F}_A$. The catheter segments following the tip adopt the previous configurations sequentially.
• Figure 2: The environment is represented by an aortic anatomy, the obstacle space $C_{obst}$, the free space $C_{free}$, the centerline space $C_{centerline}$, and the target space $C_{target}$. The catheter moves from the start configuration $\bm{q_0}$ and proceeds to move to reach the target configuration $\bm{q_g}$. (A) Top view on the aortic model. (B) Cross-sectional view of the open lumen of the descending aorta.
• Figure 3: The proposed cgail network architecture. The extrinsic reward signal considers the reward given by interacting with the environment, such as curriculum and PPO modules. The intrinsic reward has a policy that considers other factors, and it is defined inside the learning algorithm: for gail about the similarity of the path with respect to the expert demonstrations, for curiosity about the difference between the predicted and the actual path.
• Figure 4: Workflow for simulation and in-vitro user study. First, a 3D model is reconstructed from CTA images for a specific patient, and a deformable environment is built. Next, based on expert demonstrations, the C-GAIL network is trained to provide an optimal path. This path is then rendered through a GUI for the in-vitro experiments and serves as path guidance for users.
• Figure 5: In-vitro experimental setup to validate the proposed path planning approach: (1) silicone aortic phantom; (2) robotic catheter; (3) Aurora EM field generator; (4) sleeve-based catheter driver; (5) wireless controller; (6) the GUI as visual feedback.