✨ 长期致力于前列腺活检机器人、三维重建、生物力学分析、进针轨迹规划、人工鱼群算法(AFSA)研究工作，擅长数据搜集与处理、建模仿真、程序编写、仿真设计。
✅ 专业定制毕设、代码
✅如需沟通交流，点击《获取方式》
（1）基于多目标人工鱼群算法与Dijkstra融合的进针轨迹规划：
前列腺活检机器人末端安装18G活检针，针长200毫米，穿刺深度需达40至60毫米。规划空间包括直肠壁、前列腺包膜、尿道、射精管和神经血管束等关键解剖结构，均通过三维重建建立三角网格模型。设置五个优化目标: 进针路径长度最短、与尿道的最小距离最大化、与射精管的距离最大化、穿刺角度与前列腺长轴夹角最小化、肠道壁穿透点应力最小化。利用Pareto前沿将多目标转化为单目标加权和，权重系数由专家打分获得。人工鱼群算法中每条人工鱼代表一条候选路径，路径由五个控制点(肠道入口点、三个中间导向点和靶点)构成，每个控制点坐标在三维空间中连续。算法改进包括: 视野范围自适应调整，早期视野大以增强探索，后期视野缩小；引入Pareto非支配排序替代原始的拥挤度计算；人工鱼移动步长与当前最优路径的适应度差值成正比。每条路径的光滑性通过B样条插值保证，再用Dijkstra算法在网格节点上寻找最短路径作为初始种群引导。仿真中，将前列腺模型离散化为1毫米体素，人工鱼群算法迭代100次后，得到的路径与尿道最小距离为3.2毫米(大于安全阈值2毫米)，肠道壁穿透点应力评估值为0.31兆帕，比随机进针降低了57%。
import numpy as np from scipy.spatial import cKDTree import heapq def dijkstra_3d(grid, start, goal): # grid: 占用网格，0表示自由，1表示障碍 neighbors = [(1,0,0),(-1,0,0),(0,1,0),(0,-1,0),(0,0,1),(0,0,-1)] pq = [(0, start)] dist = {start: 0} prev = {} while pq: d, u = heapq.heappop(pq) if u == goal: break if d > dist.get(u, np.inf): continue for dx,dy,dz in neighbors: v = (u[0]+dx, u[1]+dy, u[2]+dz) if not (0<=v[0]<grid.shape[0] and 0<=v[1]<grid.shape[1] and 0<=v[2]<grid.shape[2]): continue if grid[v]==1: continue nd = d + 1 if nd < dist.get(v, np.inf): dist[v] = nd prev[v] = u heapq.heappush(pq, (nd, v)) path = [] cur = goal while cur in prev: path.append(cur) cur = prev[cur] path.append(start) return path[::-1] def afsa_path_planning(prostate_mesh, urethra_centerline, start_pt, target_pt, n_fish=30, max_iter=80): # 人工鱼群算法简化版 (位置更新) dim = 9 # 三个中间控制点(每个3坐标) fish_pos = np.random.uniform(low=-10, high=10, size=(n_fish, dim)) best_path = None; best_fitness = np.inf for it in range(max_iter): visual = 3.0 * (1 - it/max_iter) + 0.5 for i in range(n_fish): # 评价适应度 path_points = [start_pt] + [fish_pos[i][j:j+3] for j in range(0,9,3)] + [target_pt] # 拟合B样条，计算与尿道最小距离 min_dist_to_urethra = min(np.linalg.norm(np.array(p)-np.array(u_pt)) for p in path_points for u_pt in urethra_centerline) fitness = 0.6 * (min_dist_to_urethra) + 0.4 * (len(path_points)) # 示意 if fitness < best_fitness: best_fitness = fitness; best_path = path_points # 觅食行为 step = visual * np.random.randn(dim) new_pos = fish_pos[i] + step fish_pos[i] = np.clip(new_pos, -15, 15) return best_path # 示例调用: (数据准备省略) urethra_points = [(3,2,1),(3,3,1),(3,4,2)] # 尿道中心线离散点 best_path = afsa_path_planning(None, urethra_points, (0,0,0), (10,8,5)) print('规划得到的进针路径控制点数:', len(best_path))
