Remove `speech_monitor` dependency from `asr_worker` service definition.

src/voice_dialogue/api/core/service_factories.py

@@ -103,7 +103,7 @@ def get_core_voice_service_definitions(system_language: str, tts_config: BaseTTS
         ServiceDefinition(
             name="asr_worker",
             factory=lambda: ServiceFactories.create_asr_worker(system_language),
-            dependencies=["speech_monitor"]
+            dependencies=[]
         ),
 
         # LLM文本生成服务