diff --git a/PointAgents.py b/PointAgents.py
new file mode 100644
index 0000000..d60fd51
--- /dev/null
+++ b/PointAgents.py
@@ -0,0 +1,41 @@
+import numpy as np
+
+class PointAgent():
+    import collections
+
+    def __init__(self, goal, maxdist):
+        self.goal = goal
+        self.loc = [0, 0]
+        self.maxdist = maxdist
+
+    def dist_to_goal(self, loc):
+        return ((loc[0] - self.goal[0]) ** 2 + (loc[1] - self.goal[1]) ** 2) ** 0.5
+
+    def get_action_distr(self, loc):
+        dist = min(self.maxdist, self.dist_to_goal(loc))
+        angle = np.arctan((loc[1] - self.goal[1]) / (loc[0] - self.goal[0]))
+        return (dist, angle)
+    
+    def move(self, dist_distr, angle_distr):
+        distance = dist_distr()
+        angle = angle_distr()
+        self.loc[0] += np.cos(angle) * distance
+        self.loc[1] += np.sin(angle) * distance
+
+class NoisyPointAgent(PointAgent):
+    def __init__(self, goal, maxdist, dist_noise, angle_noise):
+        self.dist_noise = dist_noise
+        self.angle_noise = angle_noise
+        PointAgent.__init__(self, goal, maxdist)
+        
+    def get_action_distr(self, loc):
+        no_noise = list(PointAgent.get_action_distr(self, loc))
+        print(no_noise)
+        noisy = no_noise
+        noisy[0] *= self.dist_noise()
+        noisy[1] += self.angle_noise()
+        return noisy
+
+if __name__ == '__main__':
+    agent = NoisyPointAgent(goal=(10,10), maxdist=5, dist_noise = lambda: np.random.normal(loc=1), angle_noise = lambda: np.random.normal(scale=0.1))
+    print(agent.get_action_distr(loc=(0,0)))